Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Von der Fakultt fr Mathematik, Informatik und Naturwissenschaften der RheinischWestflischen Technischen Hochschule Aachen zur Erlangung des akademischen Grades eines Doktors der Naturwissenschaften genehmigte Dissertation
Berichter: Universittsprofessor Dr. Dr. h.c. Bernhard Blmich Universittsprofessor Dr. -Ing. Edmund Haberstroh
Tag der mndlichen Prfung: 12 December 2007 Diese Dissertation ist auf den Internetseiten der Hochschulbibliothek online verfgbar.
Contents
Contents
1 2 INTRODUCTION AND MOTIVATION
2.1 Polyethylene 2.1.1 History 2.1.2 2.1.3 2.2 2.2.1 2.2.2 2.2.3 Properties and economic relevance Morphology of polyethylene History Properties and economic relevance Morphology of isotactic polypropylene
1
5 5 5 5 7 7 8 8
Polypropylene
EXPERIMENTAL METHODS
3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 3.10 Introduction Transmission electron microscopy Scanning electron microscopy Atomic force microscopy Differential scanning calorimetry Small-angle X-ray scattering Wide angle X-ray diffraction Infrared spectroscopy Mechanical tests Nuclear magnetic resonance 3.10.1 Proton wide-line NMR spectroscopy 3.10.2 Proton transverse magnetization relaxation 3.10.3 Spin-diffusion NMR experiments with dipolar filters
13
13 13 13 14 14 14 15 15 16 17 17 18 19
27
27 30 30 31
Contents 4.5 4.6 4.7 4.8 4.9 Temperature dependence of the phase composition and chain mobilities of HDPE by NMR The effect of annealing on chain mobility and the phase composition by NMR The thickness of domains in HDPE The effect of annealing on the thickness of domains and chain mobility Conclusions 35 36 39 40 32
ISOTACTIC POLYPROPYLENE
5.1 5.2 5.3 5.4 5.5 5.6 Introduction and motivation Materials Crystallinity by DSC Morphology of iPP by TEM Crystallinity and Morphology by SAXS Solid-state NMR study of phase composition, chain mobility, and domain thickness 5.6.1 5.6.2 5.6.3 5.6.4 5.6.5 5.6.6 5.6.7 5.6.8 5.7 The effect of annealing temperature and annealing time on the phase composition and chain mobility Comparison of the amounts of the rigid fraction/crystallinity obtained by NMR, SAXS, and DSC Comparison of the domain thicknesses measured by DQ and Goldman-Shen dipolar filters in spin-diffusion experiments Temperature dependence of the domain thickness The effect of annealing temperature and annealing time on the domain thickness of iPP sample Thickening of crystalline domains during annealing Correlation between the 1H transverse magnetization rate and the domain thickness of the crystalline domains Conclusions
43
43 44 44 47 48 51
65
65 66
Contents 6.3 6.4 6.5 6.6 6.7 6.8 6.9 6.10 NMR data processing Mechanical results X-ray results DSC results Changes induced by aging at 28 C in the phase composition and chain mobility of homopolymeric iPP samples Changes induced by aging at 70-130 C in the phase composition and chain mobility of homopolymeric iPP samples Discussion: physical ageing in homopolymeric iPP samples by NMR and mechanical analysis Conclusions 79 80 75 73 67 67 69 69
103
103 104 105
Contents 8.4 8.5 8.6 8.7 8.8 8.9 Crystallization temperature and heat of fusion by DSC Tensile and flexural tests Phase content and chain mobility by NMR at 70 C Aging in impact modified PP copolymer at 70 C by NMR Phase content and chain mobility by NMR at 0 C Conclusions 108 109 111 112 114 118
119 123
Abbreviations and Symbols SAXS SEM SEPS SPE ta td T1 T2 Tg Tm0 TEM XRD W WAXD WAXS WLF wt% 1/2 small angle X-ray scattering scanning electron microscopy solid echo pulse sequence single pulse excitation aging time mixing time longitudinal relaxation time transverse relaxation time glass transition temperature equilibrium melting point transmission electron microscopy X-ray diffraction crystallinity by DSC wide angle X-ray diffraction wide angle X-ray scattering William Landel Ferry equation weight percent full line width at half height proton density correlation function draw ratio
Chapter 1
Introduction and Motivation
The extensive and still increasing usage of polymeric materials stems from their unique physical and mechanical properties. Combined with economic advantages in terms of price and market availability polymeric materials became a logical choice for a wide range of applications. For many purposes, their performance is much better than that of conventional materials such as metals, ceramics or wood. Among the polymeric materials, polyolefins hold an important role, becoming more and more an indispensable part of our daily life. We can identify so many products around us ranging from our basic necessities such as tooth brushes, clothing, storage bottles, and carry bags to special applications like gas and water pipelines, automotive applications, and biomedical implants all made from polyolefins. The macroscopic properties of polyolefin materials are not only determined by the size and construction of the application but to a great extend also to the morphology of the polyolefin used. Polyolefins show a very complex morphology in which various levels of hierarchy are distinguished. Four levels of morphology can be identified: chain characteristic (primary level), crystal unit-cell (secondary level), lamellae structure (tertiary level), and crystal aggregates or super-molecular structure (quaternary level). A wide range of material properties are achieved depending on the chain characteristics, type of polyolefin monomers, e.g. homo- or co-polymer, their architecture, e.g. linear, branched, microstructure, e.g. tacticity, amount and distribution of region- and/or stereo defects and molecular weight and molecular weight distribution. By incorporation of olefin comonomers macroscopic properties such as toughness and environmental crack resistance can be influenced. The morphology of the polyolefin on the tertiary and quaternary levels not only is influenced by the chain characteristics but to a large extend also by the processing conditions, storage conditions, and influence of additives. Due to this variety of possibilities in morphology and with that in the properties, the study of the structure - properties relationships in polyolefins has become an exciting field of research in recent decades. Different NMR methods applied to polyolefins can provide structural and dynamical information about their morphology that subsequently can be correlated with the macroscopic properties. The potential and the wide range of applicability of 1H solid - state NMR is emphasized in this work, which focuses on the applications of NMR in polyolefins 1
Chapter 1. Introduction and Motivation characterization. Classical and advanced solid-state NMR methods were successfully applied for the investigation of a broad range of systems such as, high-density polyethylene (HDPE), isotactic polypropylene (iPP) and PP impact modified copolymers (ICP). A good understanding of their microscopic structure and dynamics as well as morphology is necessary in order to manufacture materials with improved macroscopic properties. The primary aim of this work is to investigate the changes in the morphology, e.g. phase composition, chain mobility, and domain sizes of polyethylene and polypropylene samples as a function of composition, (comonomer incorporation, chain structure etc.), storage time, temperature at storage, speed of cooling down from the melt, and behaviour under loading conditions. Another aim relates to investigations of the structure property relationship in polyethylenes and polypropylenes. For these purposes, several experimental methods and methods of data analysis were used, as detailed in Chapter 3 wherein a comprehensive description of the 1H solid state NMR methods utilized in this work is given. The changes in the morphology, phase content, chain mobility, and domain sizes of HDPE samples as a function of annealing temperatures and storage time are presented in Chapter 4. In this chapter, we established the temperature range for which the largest differences are observed in molecular mobility in rigid, semi-rigid, and soft fractions allowing accurate determination of the phase composition and the thickness of these domains. The domain thickness, which was determined by NMR, is in good agreement with those measured by small-angle X-ray scattering (SAXS) and transmission electron microscopy (TEM). It is shown that annealing is accompanied by structural reorganizations in the amorphous layer adjacent to the lamella surface causing a continuous shift of the interface towards the inner part of the amorphous regions and thus reducing the thickness of the amorphous layer. The temperature dependence of the spin diffusivities is reported for the three phases of HDPE. The spin-diffusion experiments performed by time-domain low-field NMR and frequency-domain high-field NMR are compared. Chapter 5 focuses on the combination of different experimental techniques, e.g. DSC, SAXS, WAXD,
1
determination of the amount of rigid / crystalline, semi-rigid, and soft fractions of isotactic polypropylene (iPP). Changes in the rigid, semi-rigid, and soft fractions of iPP were investigated as a function of annealing time and annealing temperature. The most probable iPP morphology was established by TEM in combination with results from1H spin-diffusion data compared to data from multidimensional solutions of the spin-diffusion equations. The 2
Chapter 1. Introduction and Motivation temperature dependence of spin diffusivities was taken into account, and a semiquantitative theory is presented for this dependence in the case of the amorphous domains. A combination of 1H spin-diffusion, NMR and SAXS was used to estimate the lamellar thicknesses for non-annealed and annealed iPP samples. Annealing at temperatures above 110o C causes increases in the lamellar thickness and the crystallinity, and a decrease in the chain mobility of the rigid and semi-rigid fractions. The quantities of the different phases and the chain dynamics are reported for three annealing temperatures 134 C, 143 C, and 153 C, and annealing times in the range from 15 minutes to 30 hours. Chapter 6 focuses on the changes in the phase composition and chain mobility in injection moulded, melt-crystallized and quenched homopolymeric iPP samples, associated with aging at temperatures well above Tg for 150 and 1000 hours. Those changes are studied with high sensitivity using low field 1H solid-state NMR and X-ray diffraction (XRD). By NMR the aging effect at 28 C was monitored each hour during 1000 hours for all iPP samples. Aging in the temperature range from 70o C to 130o C was studied each hour during 150 hours for melt-crystallized homopolymer iPP samples. The aging effects of the morphology of the homopolymer iPP are rationalized in terms of increased restrictions of the chain mobility in the semi-rigid and soft fractions. All sample exhibit physical aging when raised to elevated temperatures, and the physical aging kinetics was observed to depend on the morphology of the homopolymeric iPP and aging temperatures. A significant increase in the tensile modulus with time was observed for injection molded homopolymeric iPP samples. The observed property changes induced by aging are attributed to micro structural changes within the semi-rigid and amorphous phases. This is consistent with a reduction in chain mobility of the semi-rigid and soft fractions. In chapter 7 changes in the amount of the rigid, semi-rigid, and soft fractions, molecular mobility, and domain thickness of uniaxially stretched isotactic polypropylene (iPP) were investigated as a function of temperature, draw ratio, drawing temperature, and drawing rate, using 1H solid - state NMR at low field. The amount of the rigid fraction determined by NMR, is in good agreement with crystallinity measured by IR. The study largely focuses on strain-induced changes in the soft phase that have not been analyzed previously in detail. The effect of drawing temperature on the studied molecular parameters shows an important effect of the rate of molecular motions on the strain - induced transformation of the spherulitic morphology of iPP to the fibrillar one. Chapter 8 presents the influence of the rubber content and nucleating agents on the morphology, phase content, chain mobility, and mechanical properties of impact 3
Chapter 1. Introduction and Motivation copolymeric PP samples. Changes in the morphology, phase content, and chain mobility were studied using a combination of different techniques like NMR, X-Ray, TEM, SEM, AFM, and DSC. Structure-property relationships for impact modified copolymeric PPs are also presented in this chapter.
Chapter 2
Polyethylene and Polypropylene Morphology
2.1 Polyethylene
2.1.1 History
In 1953 Karl Ziegler discovered heterogeneous catalyst based on titanium halides (TiCl31/3 AlCl3) that produced HDPE upon activation with an organoaluminium cocatalysit such as Al(C2H5)2Cl by coordination polymerization at low temperature and pressure [Nat1955]. Later, Natta independently produced stereoregular polymers such as propylene, butane - 1 and styrene by the same principle [Nat1955]. The subsequent discovery of a MgCl2 supported TiCl4 catalyst system resulted in a two fold increase in the activity [Nat1955, Nat1956, Nat1964].
a)
b)
c)
Figure 2.1 Chemical structures of various kinds of polyethylene (a) HDPE (linear) (b) LLDPE (many equal short branches) (c) LDPE (various branches on branches)
b a
Figure 2.2 Crystal structure of polyethylene showing a view along the chain axis: orthorhombic unit cell.
The monoclinic phase of polyethylene The monoclinic cell was first discovered by Teare and Holmes [Tea1957]. The monoclinic cell is less stable than the orthorhombic cell and is found in samples subjected to mechanical 6
Chapter 2. Polyethylene and Polypropylene Morphology stress (Fig. 2.3). The dimension at 23 C of the monoclinic cell of linear polyethylene with 4 CH2 groups are according to Seto et al. [Set1968]; a = 8.09, b = 2.53 (chain axis), and c = 4.79, which gives a crystal density of 997 kg m3. The zigzag planes of the chains of the monoclinic cell have a uniform orientation (Fig. 2.3).
a
Figure 2.3 Crystal structure of polyethylene showing a view along the chain axis: monoclinic unit cell.
The hexagonal phase of polyethylene The high - pressure hexagonal phase was discovered by Basset et al. [Bas1974]. The hexagonal phase has very special properties: a regular hexagonal packing of conformational disordered chains with extreme translational chain mobility that allows rapid crystal thickening.
2.2 Polypropylene
2.2.1 History
Polypropylene (PP) was discovered in the early 1950s. In 1953 Ziegler discovered that polyethylene could be prepared with a mixture of metal salts and transition metal salts [Bus2001, Cip2001]. Giulio Natta succeeded in preparing polypropylene using the Ziegler catalyst and was able to obtain and characterize isotactic PP by fractionation [Sal1996]. In 1957 polypropylene was taken into commercial production [Kar1995]. The polymer yield of the catalyst (so-called Ziegler - Natta catalysts) had increased enormously from 0.8 Kg/(gram catalyst) to more than 100 Kg/(gram catalyst) along with a significant increase in isotacticity. In normal practice the term polypropylene means isotactic polypropylene obtained from Ziegler - Natta catalysis.
b.
c.
d.
Figure 2.4 Isomerism in polypropylene. a) Chemical structure of polypropylene. b) Isotactic polypropylene. c) Syndiotactic polypropylene. d) Atactic polypropylene, as illustrated by Karger Kocsis [Kar1995]
The three physical stereo - configurations that can be distinguished in polypropylene are: isotactic, syndiotactic and atactic (as presented in Fig. 2.4b, c, d, respectively). From 13C NMR it is possible to extract the information about the degree of regularity within the polymer.
R up L dw R dw L up a sin
Figure 2.5 Two limiting forms of the - phase. Left: 1; Right 2. The horizontal arrows indicate the projection of the a axis, the vertical arrow indicates the b axis. The c axis is perpendicular to the plane of view [Aur2000].
One of the characteristic morphological features of the - phase is the so called cross hatching, also known as lamellar branching. This feature is recognized by the initial lamella (mother lamella) showing branches (daughter lamellae) that are attached to mother lamellae with angles of 80 or 100 . The molecular background of crosshatching is described by Lotz and Wittmann [Lot1996]. The - phase of polypropylene Under proper crystallization conditions (shear, large temperature gradient, or use of a nucleating agent), polypropylene samples with a high content of - phase can be made [Vle1997, Tri1998, Per1999].
L L
Figure 2.6 Schematic representation of the arrangement of iPP helices in the - phase. The arrows indicates the b axis, the c axis is perpendicular to the plane of view.
The - phase has a trigonal unit cell with lattice parameters a = b =11, c = 6.5, = 120. The - phase is metastabile relative to the - phase.
10
Chapter 2. Polyethylene and Polypropylene Morphology The phase of polypropylene The phase is usually associated with the phase, although individual single crystals of the phase has also been observed. The modification may be formed in degraded, low molecular weight isotactic polypropylene and in the samples crystallized under high pressures [Mez1997, Cam 1993]. Also, considerable amounts of the modification were obtained under atmospheric pressure, both in systems with low tacticity or made by homogeneous metallocene catalysts [Per1999, Mez1998]. The crystalline structure of the modification was first denoted as triclinic [Per1999, Lov1983]. However, it has been shown that the structure is orthorhombic with parameters a = 8.54, b = 9.93 , and c = 42.41 [Cam1993]. The structure is composed of sheets of parallel chains in which the molecular orientations in adjacent sheets are inclined at an angle of 80 [Tho1996]. In this unit cell the c axis is not parallel to the chain axis direction.
Figure 2.7 Schematic representation of the arrangement of iPP helices in the phase [Cam1993, Lot1991].
11
12
Chapter 3
Experimental Methods
3.1 Introduction
In this thesis PE and PP samples are investigated using transmission electron microscopy (TEM), atomic force microscopy (AFM), scanning electron microscopy (SEM), differential scanning calorimetry (DSC), X-Ray diffraction (XRD) and nuclear magnetic resonance (NMR). All these experimental methods and the associated experimental issues are described in brief in this chapter.
13
14
2d =
n , sin
where n is an integer. The equation indicates a reciprocal relationship between the characteristic length d and sin .
15
where: Ec is the corrected energy, in joules, absorbed by breaking the specimen h is the thickness of the specimen in mm bN is the width at the notch in mm Two hammers were used having a kinetic energy of 2.75 J and 5.5 J, respectively. If after carrying out the test with the lighter hammer, the specimen did not fully brake, the specimen was discarded without recording the impact strength and the heavier hammer was used. Flexural testing method: The principle of this test is to bend or flex the specimen in the middle at a constant bending rate until the specimen fractures or until the deformation reaches some predetermined value. During the test the force applied to the specimen is measured. Flexural modulus was computed and used as an indication of the materials stiffness when flexed. The tests were carried out with a Zwick tensile machine according to ASTM D790. The dimensions of the specimen were 65 x 12 x 3.2 mm. The values of flexural modulus are an average values of five tests. Tensile testing method: Is used to investigate the tensile behaviour of the specimens and for determining the tensile strength and tensile modulus. The tests were carried out with a Zwick Z050 tensile machine according to ISO 37-II. The tests were performed at 25 C with a deformation rate of 50 mm/min. The aim of the tensile tests were to determine the effect of the rubber content and the addition of the nucleating agents on the yield point of the specimens under investigation.
16
Figure 3.1. (a) Low field and (b) high field NMR spectrometers [Web1, Web2]
17
Figure 3.2. Proton wide-line NMR spectra (solid line) of iPP measured at 70 C with a 500 MHz NMR spectrometer. The dashed lines show the spectral components that are assigned to the rigid, semi-rigid, and soft fractions of iPP corresponding to broad, intermediate, and narrow line widths, respectively.
transverse magnetization as a function of time t. The second experiment was the solid echo
pulse sequence (SEPS), i.e. 90x tse 90y tse acquisition of the amplitude of the
transverse magnetization A(t), with tse=10 s. The third one was the Hahn-echo pulse
sequence (HEPS), i.e. 90x tHe 180y tHe acquisition of the amplitude of the echo
18
A(t)/A(0)
Hahn-Echo 0.1 soft fraction rigid semi-rigid fraction fraction FID 0 100 200 300 time [s]
Figure 3.3. The decay of the transverse magnetization relaxation (points) for iPP at 70 C. The decay (FID) was measured at 20 MHz NMR spectrometer using the SPE ( ) and the HEPS (I) methods. The solid line represents the result of a least-squares fit of the decay with a linear combination of the Abragam function and two exponential functions. Dotted lines show the separate components that are assigned to rigid, semi-rigid, and soft iPP fractions.
The Hahn-echo pulse sequence was used to avoid the effect of the inhomogeneity of the magnetic field on the decay of the transverse magnetization of the soft fraction of iPP, and to obtain reliable information about the molecular mobility and amount of the soft fraction. The 90o pulse excitation (SPE) and HEPS were finally used to record the FID for quantitative analysis of the phase composition (Fig. 3.3). Repeated experiments for the same sample indicated that the relative error of the extracted relaxation parameters was about 1 %.
19
diffusion time td . The evolution time of the DQ coherences is t DQ which was taken as 5 s in all experiments.
3 DQ signal [arb.units]
0 0
50
100 [s]
150
200
Figure 3.4. Proton DQ build-up curve for iPP at room temperature showing the dependence of the integral signal on the excitation time . The DQ build-up curve was measured with high-field NMR spectrometer. The maximum of the curve is observed at 10 s that is marked by a dashed line.
The dipolar filter excites double-quantum coherences and selects mainly the magnetization of the rigid fraction at a short excitation time [Bud2003, Hed2007]. The optimum value in the DQ filter for selection the rigid fraction [Sch1999] was obtained from the DQ build-up curve. It shows one maximum for iPP (Fig. 3.4) at short excitation times of approximately 10 s. The efficiency of the filter is judged from 1H wide-line NMR spectra recorded at different excitation/reconversion times (Fig. 3.5). These spectra allow us to choose the optimum filter time for selecting the magnetization of the rigid fraction. For short excitation/reconversion time , the DQ filtered spectra are doublets (Figs. 3.5a and 3.5b) which are related to the spin pairs of the methylene groups having the strongest dipolar couplings. It should further be mentioned that the doublet structure of the DQ-filtered spectra is due to the fact that at short excitation times, only the rigid polymer chains with dipolar tensor orientation around 0 and 90 degrees are preferentially excited, while those orientated around the magic angle are suppressed. An interesting property of the DQ filter is that at long excitation times, only the signal from the most mobile chain fragments in the amorphous phase is selected (Fig. 3.5d).
20
b)
=6 s
-400
c)
-200
0 ppm
200
400
d)
-400
-200
0 ppm
200
400
=20 s
= 40 s
-400
-200
0 ppm
200
400
-400
-200
0 ppm
200
400
Figure 3.5. High-field proton wide-line NMR spectra of iPP at room temperature recorded after different excitation/reconversion periods of the DQ dipolar filter: = 2 s (a), = 6 s (b), = 20 s (c), and = 40 s (d). The spin-diffusion time td was set to 5 s.
The DQ filter shows high efficiency in selecting the magnetization of particular iPP and PE domains with distinct molecular mobilities. The high selectivity of the DQ filter was also observed in T2 relaxation experiments at low resonance frequency. For all experiments, the excitation/reconversion time of 10 s and short spin-diffusion time td of 5 s (Fig. 3.6) were chosen to select the rigid iPP and PE fractions in the spin-diffusion experiment, both at high- and low- proton resonance frequencies. The proton NMR spectra edited for different spin diffusion times are shown in Fig. 3.6.
21
td = 5 s x0.57
td = 10 s x0.71
-200
c)
-100
0 ppm
100
200
d)
-200
-100
0 ppm
100
200
td = 10 ms x0.90
td = 1 s x1
-200
-100
0 ppm
100
200
-200
-100
0 ppm
100
200
Figure 3.6. Proton wide-line NMR spectra of iPP at room temperature. The spectra were recorded using the spin-diffusion experiment with different spin-diffusion times: a) td = 0.005 ms, b) td = 0.010 ms, c) td = 10 ms and d) td = 1 s, and =10 s.
Goldman Shen filter. Another dipolar filter used in this investigation explores the GoldmanShen scheme [Gol1961]. The amorphous domains are selected by the pulse sequence
90 0 90 0 x t d 90 0 FID with the filter time = 100 s and td being the spin-diffusion x x time. Both dipolar filters were employed in the present study to gain a higher confidence in determinations of the polyolefins morphology. Spin-diffusion coefficients. The values of the spin-diffusion coefficients should change with increasing temperature due to an increase in molecular mobility. Therefore, the spindiffusion coefficients should be determined at different temperatures for accurate analysis of domain sizes in PE and PP as a function of temperature. The values of the spin-diffusion coefficients D for the rigid and the soft fractions of PE and PP can be determined assuming, to a good approximation, that the NMR line shapes of the rigid and soft fractions are Gaussian and Lorentzian, respectively. The spin-diffusion coefficients can be related to the second van Vleck moment of the NMR absorption lines,
22
which, in turn, is related to the full line width 1/2 at half height. Hence, the spin diffusivity for the rigid (Dr) and soft (Da) fractions is as follows [Dem1995]:
Dr
1 r 2 1 / 2 , 12 2 ln 2
1 2 1/ 2 r [ 1 / 2 ] , 6
(3.1)
and Da (3.2)
where r 2 is the mean square distance between the nearest spins, and is a cut-off parameter for the Lorentzian line shape [Dem1995]. The temperature dependence of 1/2 is determined at each temperature by spectral deconvolution as shown at one temperature in Fig. 3.2. For calculating r 2 the Discover program (version 2004.1) was used. The estimated weighted
mean square of these distances r 2 equal approximately 0.0484 nm2 for PE and 0.08 nm2 for PP, respectively. The calculated spin-diffusion coefficients are plotted as a function of temperature in Fig. 3.7. The temperature dependence of the spin diffusivity in different fractions of iPP is difficult to predict quantitatively, due to the complex origin of the chain motion and multispin interactions. Nevertheless, a semi-quantitative approach can be presented based on the arguments provided in Ref. [Abr1961]. In the presence of motions of the spins, the theory of the NMR line width predicts that 1 / 2 M 2,rigid c , where M 2,rigid is the second van Vleck moment for a rigid polymer, and c is the correlation time of the chain motions.
40
60
temperature [ C]
Figure 3.7. The temperature dependence of the spin-diffusion coefficient for the rigid Dr and amorphous Da fractions of iPP.
23
From Equation 3.2 and the above relationship, one obtains: Da ( c ) Williams-Landel-Ferry (WLF) equation (see Ref. [Sch1994]), i.e.,
1/ 2
. The temperature
c (T ) = c (Tg )exp
(3.3)
where Tg is the glass transition temperature and C1 and C 2 are the WLF coefficients. Using the above equation, the temperature dependence of the spin diffusivity Da of the amorphous fraction of iPP can be described by the following equation.
C1 2 (T Tg ) , Da (T ) = Dag exp T (Tg C 2 )
(3.4)
where Dag is the spin-diffusion coefficient at temperature Tg . The spin-diffusion coefficient of the rigid fraction at 25 C is taken as Da 0 Dag (see Fig. 3.7). A least-squares fit of the dependence in Fig. 3.8 provides the following values of the WLF
log[Da(T)/Dag(T0)]
T-Tg [ C]
Figure 3.8. The dependence of log[Da(T)/Dag(T0)] on temperature. The line shows the fit of the dependence with Eq. (3.4).
These values are in the range of typical WLF coefficients that describe dynamic mechanical and dielectric data for amorphous polymers, namely
C1 16
and
C 2 = 25 K 104 K [Str1997]. The observed differences could be due to the fact, that the WLF
24
function has originally been proposed to describe the temperature dependence of the viscosity. An increase in the amplitude and the length scale of the chain motion with increasing temperature could affect the spin-diffusion coefficients and mechanical data in a slightly different way.
Morphology of polyofefins by 1H spin-diffusion. The ability of NMR spin-diffusion
experiments to provide self-consistent information on the dimensionality of the diffusion process was discussed in Ref. [Dem1996]. Proton spectra recorded at 70o C with the DQ filter after different diffusion times for iPP are shown in Fig. 3.6. The flow of longitudinal magnetization from the rigid domains to the amorphous domains is observed with increasing diffusion time, causing an increase in the intensity of the narrow line at the expense of the broad line. The time-dependent integral of the spin-diffusion spectra intensities reach their equilibrium values for iPP samples in about 70 ms. In a good approximation, the spindiffusion DQ edited 1H NMR spectra can be decomposed into two components. The twocomponent analysis is used to establish the average thickness of rigid (crystalline) and amorphous domains. It can be shown that the solutions of the spin-diffusion equations for different morphologies exhibit different sensitivities to the dimensionality of the process, as a result of different surface/volume ratio (interface area) for a given source volume. The information about the dimensionality of the spin-diffusion process can be obtained using the following equations [Vod2006]: rE = 1 , 1D case p rE = 1 , 2D case p(2 + p) rE = 1 , 3D case (1 + p)3 1 where r =
r
amorphous and rigid domains are denoted as M a , eq and M r , eq , respectively. The ratio between the domain thicknesses da and dr is denoted by p = d a / d r . In the two-phase approximation, the r and E values for non-annealed iPP samples are obtained as 1.11 and 0.25, respectively. These values are established independent of the spin-diffusion process. The spin-diffusion decay and build-up curves shown in Fig. 3.9 are fitted with the solutions of the spin-diffusion equations for 1D morphology [Bud2003,Wan1996]. For each spin-diffusion dimensionality, the values of d a and d r are established from the best fit (Table 3.1). These values are used to 25
estimate the ratio p . The estimated dimensionality ratios (Eqs. (3.5)) are near unity for the 1D lamellae morphology (Table 3.2).
5 (td)
10
1/2
15
1/2
20
[ms ]
Figure 3.9. Proton spin-diffusion decay and build-up curves for the rigid () and amorphous () fractions of iPP recorded using the DQ dipolar filter. These filters select the magnetization of the rigid fractions of iPP. The solid lines represent the best fit for the 1D solutions of the spin-diffusion data for the two-phase morphology.
Table 3.1. Long period (Lp), the thicknesses of rigid (dc) and amorphous (da) domains, at 70 C for iPP samples obtained by NMR spin-diffusion experiments for 1D, 2D and 3D morphology models. dr (nm) da (nm) Lp (nm) Model 1D 10 2.5 12.5 2D 15.4 3 18.4 3D 21 3.5 24.5 Table 3.2. The spin-diffusion dimensionality, as determined by Equation (6) for iPP. dimensionality dimensionality ratio values of the dimensionality ratio 1D rE/p 1.2 2D rE/[p(2+p)] 0.68 3D rE/[(1+p)3-1] 0.52 The values of the proton densities for the different phases of iPP, which are required for the calculation of domain thickness, are determined from the densities of the crystalline and amorphous phases of iPP, i.e., c = 0.94 g/cm3 and a = 0.84 g/cm3 [Bra1999]. No correction of the spin-diffusion data due to the longitudinal relaxation effect, has been performed, because the spin-diffusion process is nearly completed at the longest mixing time of td = 70 ms. This time is significantly shorter than T1 = 120 ms. Moreover, the absolute values of relative fractions instead of the absolute values of the signal intensities, compensate to a large extend for the effect of T1 relaxation in the present case. 26
Chapter 4
High Density Polyethylene (HDPE)
4.1 Introduction and motivation
The morphology of semi-crystalline polymers has important effects on the materials properties including the mechanical performance [Pop1987, Man1988]. A quantitative characterization of the phase composition and molecular mobility in semi-crystalline polymers is therefore of great importance to advance our understanding of their properties. In this respect, the phase composition is probably one of the most important morphological parameters, mainly because the amorphous and crystalline phases exhibit vastly different behavior and their relative contributions to the material properties should be accurately known. Traditionally, a two-phase model is used to describe the morphology of undeformed, melt-crystallized polyethylene (PE) and other semi-crystalline polymers [Flo1978, Man1990]. The most common methods for crystallinity determination are X-ray diffraction, density measurement and differential scanning calorimetry. In general, different methods for crystallinity determination do not always yield the same results on exactly the same sample [Isa1999, Hir1990] because of the following reasons: (1) The complex morphology of semi-crystalline polymers requires different sets of assumptions for the analysis of the data recorded by different techniques. (2) The discrimination of the crystalline and amorphous phases is made on the basis of different characteristics, such as the enthalpy of melting (DSC), long range periodicity (WAXD) and the specific volume (density analysis). (3) The two-phase model is a simplified for description of semi-crystalline polymers due to the presence of a crystal-amorphous interface, which can be detected either as crystalline or amorphous fraction depending on the method used [Isa1999]. Various experimental methods, such as neutron scattering, dielectric relaxation, calorimetry and solid-state NMR, show that a thin layer separates crystalline and amorphous phases, and the properties of this layer are intermediate between those for crystalline and amorphous phases [Man1988, Man1990, Iwa2002, His1999]. Therefore, the term phase composition is perhaps more appropriate than simply crystallinity to emphasize the multi-phase nature of semi-crystalline polymers. Solid-state NMR is one of the most informative techniques for characterization of molecular mobility and molecular scale heterogeneity in materials [Sch1994]. During the years, different solid-state NMR methods have been used for investigations of morphology 27
and molecular mobility in PE. At temperatures well above Tg of the amorphous phase, the T2relaxation decay and wide-line NMR spectra for PE can usually be decomposed into three components, which originate from a crystalline phase, a semi-rigid crystalline-amorphous interphase, and a soft fraction of the amorphous phase. The intermediate phase has distinct dynamic properties and may not be considered as a true thermodynamic phase. Apparently, the definition of an interface or a semi-rigid fraction of the amorphous phase is more appropriate for this phase. phase selectivity of
13 13
about molecular mobility in the different phases of PE, than 1H NMR because of the high C NMR exploring differences in the chemical shift for crystalline and amorphous phases [Lit2002, Kit1994, Hil1998, Che1994, Sch1991, Kle2002, Kuw2000]. Most of the 13C NMR studies support the three-phase model of semi-crystalline PE. However, other studies suggested four types of structures with distinctly different molecular mobilities [Hil1998, Hu2000]: two types of crystalline environments, both with an all trans chain conformation - a more perfect one and one with twist defects, and two types of chain
fragments in the amorphous phase - less mobile chain units mainly in the trans conformation
rotating around the trans chain axis, and more mobile chain fragments whose mobility approaches isotropic tumbling. Detailed information about the nature of the molecular mobility in linear PE has been obtained by two-dimensional (2D) 13C exchange NMR and 2H wide-line NMR [Sch1991, Hen1984]. Despite numerous studies, the morphological origin of PE regions with different molecular mobilities is still a matter of discussion. Therefore, the determination of the thickness of domains with distinctly different molecular mobility is of a substantial interest for a better understanding of the morphology and relaxation properties of PE. Domain sizes in heterogeneous polymers [Bri1993, Pac1984] and polymer fibers [Hu2000, Bud2003, Bud2004, Van1996] have been determined by proton NMR spin-diffusion experiments. Previous 1H NMR spin-diffusion experiments have shown that the lamellae thickness of meltcrystallized PE and PE fibers varies in a wide range depending on the degree of the branching and the thermal history of the sample [Eck1997, Pac1984]. It should be mentioned that the
thickness of the rigid fraction in semi-crystalline polymers could differ from the lamellae thickness since some fraction of the amorphous phase adjacent to the lamellae surface can be
largely immobilized and cause an apparent increase in the lamellae thickness. The estimated thickness of the mobile amorphous fraction in PE ranges from 1 nm to 5 nm [Eck1997, Hu2000, Pac1984]. The morphology and phase characteristics of HDPE were also investigated by
13
magnetization from the rigid phase [Van1996]. The dimensions of 6.5 nm, 0.8 - 1.3 nm, and 5 nm were determined for the crystalline, intermediate, and amorphous phases, respectively. 13C solid-state NMR methods offer certain advantages for the analysis of domain sizes in PE due to high phase selectivity [Axe1983, Kit1986]. The high sensitivity of proton NMR makes this method very attractive for characterization of crystallization kinetics [Kri2000], premelting behaviour [Tan1995], and quality control [Bud2003]. The main disadvantage of 1H NMR for a phase analysis in PE is a lack of high selectivity regarding the crystalline and amorphous phases, if the experiments are performed at room temperature. A discrimination of these phases by proton NMR is made on the basis of differences in molecular mobility. Despite numerous 1H NMR studies of PE, it is still not well known how annealing at elevated temperatures affects the phase composition and the thickness of the domains. To the best of our knowledge neither the temperature dependence of the domain thickness was studied by NMR so far, as well as the effect of annealing on the domain thickness. Based on the above considerations, we aim to establish a reliable method for the analysis of the phase composition and domain sizes in HDPE using high-field proton wideline NMR and low-field
1
comparing the results of these two methods the performance of low-field NMR measurements for characterization of the domain sizes in semi-crystalline polymers can be validated. Furthermore, the effects of measurement temperature and annealing at elevated temperatures on the polymer morphology are studied in detail, as temperature may play a decisive role in the outcome of the measurements. Knowledge of the temperature effect will help to identity the temperature range for which the largest NMR difference between the different phases can be obtained, while ensuring high accuracy of the method by avoiding annealing. A recently developed NMR spin-diffusion experiment is applied for measuring the domains sizes [Bud2003, Bud2004]. This method explores a double-quantum dipolar filter [Bud2003, Bud2004, Ba1998, Che2005], provides higher differences, and consequently a more accurate analysis is possible. In order to evaluate the reliability of the NMR method, the results are compared to crystallinity and domain sizes determined for the same sample by SAXS and TEM.
29
4.2 Materials
High-density polyethylene (HDPE) samples were obtained from SABIC Europe BV. The molar mass and molar mass distribution, as determined by size exclusion chromatography in 1,2,4-three chlorobenzene, were characterized by: Mn = 15.103 g/mol, Mw = 78.103 g/mol, and
Mw/Mn = 5.2. Four millimetre thick compression moulded plates were prepared by melting
HDPE in a mould first under atmospheric pressure at 180 C with a subsequent pressure increase to 500 kPa. To create a homogeneous melt, the HDPE sample was kept in the mould for 15 minutes. Afterwards, the mould was cooled with cold water and the HDPE plates were annealed at 100 C in an oven for one hour followed by slow cooling to room temperature. For the annealing study, the sample was further annealed at different temperatures inside the NMR probe for 100 hours.
100 nm
Fig. 4.1 TEM picture of a core part of a compression moulded HDPE plate. The length of 100 nm is shown in the bottom left-down corner of the figure.
30
a) 3 30 C o 50 C o 70 C o 90 C
o
b) 1.0
30 C o 50 C o 70 C o 90 C
o
I(s)*s
10
20 x [nm]
30
40
Fig. 4.2 (a) One-dimensional scattering intensities I1(s)=I(s)*s2 recorded at various temperatures, indicating the maximum of the long period due to periodic stacking of crystalline lamellae. (b) Normalized linear correlation functions (x) derived from the SAXS.
The long period Lp is estimated from the position of the first interference maximum, at about 0.038 nm-1 (Fig. 4.2a), and from the normalized correlation function, which is shown in Fig. 4.2b. According to these data, the long period is close to 25 nm, which is larger than the values obtained by TEM, i.e. (18 - 19) 2 nm. It is noted that the Lp value can suffer from systematic errors due to the experimental set-up used. Based on the analysis of the correlation functions, the thickness of the crystalline domains is estimated to 17 18 nm, which is in good agreement with the value measured by TEM and NMR, as shown below. With the values of d and Lp from the correlation functions, the crystallinity is estimated to 68 5 vol%. This value corresponds to a crystallinity in mass per cent of 71 5 wt %, which is calculated with the densities of the crystalline and amorphous phases of 0.99 g/cm3 and 0.87 g/cm3, respectively. It is suggested in this 31
calculation that the density of an interfacial layer is given by the mean value of the densities for the crystalline and amorphous phases. SAXS experiments at different temperatures reveal almost the same long period, lamella thickness, and crystallinity in the temperature range from 30 C to 90 C.
4.5 Temperature dependence of the phase composition and chain mobility of HDPE by NMR
The existence of three fractions with different molecular mobilities is revealed by 1H NMR spectra, and the proton NMR free induction decay (FID) was measured for HDPE at 100 C. The relaxation components are assigned to three fractions with different molecular mobilities, the rigid - T2rigid, semi-rigid - T2semi-rigid, and soft -T2soft fractions, respectively [Ber1978, Fed1985, Han1998, Lit2004]. Like the FID, the spectra can be deconvoluted into three components (example Fig.3.3 and Fig. 3.4). Broad, intermediate and narrow lines originate from rigid, semi-rigid and soft fractions of HDPE, respectively. However, the quantitative analysis of FID and wide-line NMR spectra in terms of crystalline phase, crystal-amorphous interface and soft fraction of the amorphous phase can be complicated due to the following reasons: (1) To observe distinct differences in molecular mobility of the crystalline and amorphous phases, and consequently in the T2 relaxation and the line width, the temperature of the sample should substantially exceed the dynamic glass transition temperature at the time scale of the NMR experiment, i.e. in the range of 10-4 - 10-5 sec. The Tg of the HDPE sample that was determined by a high performance DSC equals 57o C. The NMR Tg should be observed in the temperature range from 30o C to 10o C. Therefore, a proper choice of the temperature of the NMR experiment is required. (2) Since the T2-relaxation experiments should be performed at temperatures well above Tg, the sample exposure to elevated temperatures can cause irreversible changes in the phase composition and thus in the molecular mobility. Therefore, the temperature for the experiments should not be too high to prevent annealing of the sample during the NMR experiment. In order to find the optimum temperature for determination of the phase composition by NMR, we recorded the temperature dependence of the T2 relaxation and the wide-line spectra (Fig. 4.3). At room temperature, a significant fraction of the amorphous phase is rigid on the time scale of the NMR experiments and this amorphous fraction contributes to the intensity of the broad line of the NMR spectra and the short T2-relaxation component of the FID, as can be concluded by comparing the amount of the rigid fraction with the crystallinity determined by 32
SAXS, i.e 71 5 wt%. Upon increasing temperature, the amount of the rigid fraction, which is composed of the crystalline phase and the rigid fraction of the amorphous phase, gradually decreases, whereas the amount of the semi-rigid and soft fractions increases. These changes are caused by an increase in molecular mobility of the less constrained chain fragments in the amorphous phase, and possibly by melting of thin lamellae, which are inserted into the primary stack of crystallites upon cooling of HDPE which was crystallized at elevated temperatures [Mat1994].
a)
100 80 content [%] 60 20 0 rigid semi-rigid soft
b)
100 80 content [%] 60 20 0 rigid semi-rigid soft
40
60 80 100 o temperature [ C]
120
40
60
80
o
100
120
temperature [ C]
Fig. 4.3. Temperature dependence of the amount of rigid, semi-rigid, and soft fractions of HDPE measured with (a) low-field, and (b) high-field NMR spectrometers. Dotted line shows crystallinity that was determined by SAXS in the temperature range from 30 0C to 90 0C.
In the temperature range from approximately 100 C to 120 C, the amount of rigid fraction is close to the crystallinity measured by SAXS. Above 120 C, the amount of the rigid fraction decreases due by melting. The DSC analysis of this sample shows that the extrapolated onset temperature of melting, as determined by a method described elsewhere [Cru2004] is in proximity of 125 C and the peak of melting is observed at 136 C. Thus, the temperature range from 100 C to 110 C is the most suitable for an accurate determination of the phase composition in HDPE, since (1) the largest contrast in molecular mobility in the different fractions of HDPE is observed at these temperatures, (2) no substantial sample annealing occurs as will be shown below, and (3) the amount of the rigid HDPE fraction is close to crystallinity determined by SAXS. It is mentioned in this context that, in general, different methods for determination of crystallinity do not necessarily yield the same crystallinity value for exactly the same sample.
33
100
10
40
60
80
100
120
140
temperature [C]
Fig. 4.4. Temperature dependence of 1H T2 relaxation times for HDPE. The relaxation components with short (T2rigid) intermediate (T2semi-rigid), and long (T2soft) decay time (see Figure 4.3) are assigned to the rigid, semi-rigid and soft fractions of HDPE, respectively.
As far as the molecular mobility in HDPE is concerned, the changes in the T2 relaxation times in the temperature range studied (Fig. 4.4) can be assigned to the following relaxation transitions in HDPE [Boy1985]. A small increase in T2rigid above 60 80 C is related to chain motions in the crystalline phase - the -relaxation process, which occurs prior to melting. Interpretation of the molecular mobility in the amorphous phase is more cumbersome. Despite numerous studies of relaxation processes in PE by different methods, a molecular interpretation of the relaxation processes in the amorphous phase is under discussions [Woo1993]. The increase in T2soft with increasing temperature from 40 C to 70 C can be associated with a relaxation process in the soft fraction of the amorphous phase. This temperature range and frequency of this relaxation, as estimated from change in T2soft ( 5 10 kHz), are in the range that is characteristic for the -relaxation in the amorphous phase. It is noted that the -relaxation in HDPE is very broad and hardly detectable by mechanical and dielectric relaxation methods [Woo1993]. Partial melting of HDPE crystals above 120 C causes a decrease in constraints on chain motions in the amorphous phase and therefore an increase in molecular mobility in the amorphous phase, as reflected by an increase in T2semi-rigid and T2soft (Fig 4.4).
34
4.6 The effect of annealing on molecular mobility and the phase composition by NMR
The experiments above have shown that the analysis of the phase composition should be preferably performed at elevated temperatures where the NMR parameters show enhanced contrast. Since the NMR experiments take approximately 2 hours, sample annealing is indeed an issue. In order to determine a temperature, at which no detectable changes in the relaxation will be observed during NMR experiments, the T2 relaxation was recorded in real-time at 100 C, 110 C and 121 C. The data were recorded at one-hour intervals. Sample exposure to 100 C and 110 C for 60 hours causes hardly any change in the molecular mobility and the phase composition. Annealing at 121 C for 100 hours showed an increase in the amount of the rigid fraction by 1 wt% at the expanse of semi-rigid and the soft fractions of the amorphous phase (Fig. 4.5a).
a) 78
b) 275 250 T2 rigid T2 semi-rigid T2 soft 50 25
76 content [%]
12 10
T2 [s]
225
20
40
60
80
100
0 0
20
40
60
80
time [hours]
time [hours]
Fig. 4.5. a) The amounts of crystalline phase (% T2rigid), semi-rigid crystal-amorphous interface (% T2semi-rigid) and soft fraction of the amorphous phase (% T2soft) in HDPE as a function the annealing time at 121 C. b) The effect of annealing on molecular mobility, as determined by the T2 relaxation time for the different phases of HDPE. Larger amplitude and/or frequency of molecular motions give rise to longer T2 values.
The relaxation characteristics above are determined for the crystalline phase and the crystal-amorphous interface from the analysis of the FID and using the Hahn-echo data for the soft fraction of the amorphous phase, as described in experimental part. Due to annealing both the melting temperature as well as the melting enthalpy increase, namely: the melting peak shifts from 139.1o C to 145.4o C, and the melting enthalpy increase from 232.8 J/g to 273.9 J/g. The molecular mobility in the rigid, semi-rigid, and soft fractions of HDPE is also slightly affected by annealing at 121 C (Fig. 4.5b). The value of T2 for the rigid fraction (T2rigid) 35
decreases upon annealing from 15 s to 14 s due to perfection of crystals, as can be also concluded from large increase of the melting enthalpy at only 1 wt% crystallinity increase. The T2-relaxation time for the semi-rigid fraction and soft amorphous phase decrease by a few percent. This decrease is likely to be caused by additional constraints imposed on to the amorphous phase due to the increased amount of crystalline phase. The analysis of the temperature dependence of the T2 relaxation and the annealing study show that 100 C is the optimum temperature for quantitative analysis of the phase composition in HDPE using the NMR experiments. At this temperature, the relaxation components T2rigid, T2semi-rigid, and T2soft (Fig. 4.4) largely originate from the crystalline phase, semi-rigid crystal-amorphous interface, and the soft fraction of the amorphous phase, respectively. The relative fraction of these components represent the corresponding weight fractions (Fig. 4.3a). The observation of three distinct T2 relaxations rather than a wide distribution of the relaxation times indicate a fast loss of restrictions on rotational and translational chain mobility when moving away from the crystalline phase to the interface and to the soft fraction of the amorphous phase. Therefore, a step-wise change in the molecular mobility at the position between different phases is observed. Thus, we can conclude that the three-phase model describes the phase structure of HDPE well [Woo1993, Sch1999, Pav2005].
36
0.1
0.0 0
5
1/2
(td)
1/2
[ms ]
Fig. 4.6. Proton spin-diffusion data for the rigid (), semi-rigid (), and soft fractions () of HDPE at 100 C. The experiment was performed on a low-field NMR spectrometer. The solid lines are the best fit with 1D solutions of spin-diffusion equations [Bud2003]. The data of short td are shown in the inset.
Table 4.1. The thickness of rigid (drigid), semi-rigid (dsemi-rigid) and soft domains (dsoft) at different temperatures in non-annealed HDPE. The domain thickness was obtained from 1H spin-diffusion experiments with a DQ filter.a,b The thickness of Temperature domains 70 C 100 C 120 C drigid (nm) 18.0 (19.0) 17.0 (17.0) 13 (12.5) dsemi-rigid(nm) 1.1 (1.1) 1.3 (1.4) 1.2 (1.3) dsoft (nm) 0.6 (0.5) 0.7 (0.8) 1.0 (1.1) 20.8 (21.7) 20.3 (20.6) 16.4 (16.2) Long period from NMR: d (nm) a Measured using 500 MHz and 20 MHz (in brackets) NMR spectrometers. b The relative error of the data, is estimated to about 15 %. The thickness of HDPE domains before annealing was determined using the time-domain and the frequency-domain spin-diffusion experiments, both methods yielding results that are in good agreement with each other (Table 4.1). The result for the long period is almost constant below 100o C (Fig. 4.7a). Moreover, below 100 C, the thickness of the rigid fraction drigid slightly exceeds the lamella thickness determined by TEM and SAXS. This is due to the fact that the interfacial layer with intermediate mobility on the lamellae surface is largely immobilized below 100 C, as can be seen by comparing the amount of the rigid fraction with the sample crystallinity measured by SAXS (Fig. 4.3). The thickness of the rigid domain slightly decreases with increasing temperature and at approximately 100 C reaches the lamella thickness as measured by TEM and SAXS, i.e. 16 18 nm. Thus, the thickness of rigid domain at 100 C represents lamellae thickness. The lamellae thickness at 120 C is 37
smaller than that at 100 C, which could be caused by partial melting of the sample. Moreover, fragmentation of lamellae upon melting could also cause an apparent decrease in lamellar thickness due to an increase in the efficiency of the spin-diffusion. As far as the soft domain is concerned, its thickness increases with increasing temperature at the expense of the rigid domains. The thickness of the interface is nearly constant in the studied temperature range and equals 1.1 1.3 nm (Fig.4.3c). This value is in the same range as previously estimated by Monte Carlo simulations [Gau2000, Bal2000, Man1990]. The simulations have shown that density, orientational order and transverse structure parallel to the lamella surface change from those of the crystalline phase to the disordered amorphous phase within a layer of 1.0 1.2 nm thickness that is adjacent to the lamellae surface. It should be noted that the interface thickness is close to the length of the statistical segment for PE chains, which consists of 7 7 carbon carbon bonds [Aha1983] and equals to 0.8 nm for fully extended PE chain. Therefore, distinct differences in chain mobility in different fractions of HDPE are apparently caused by short-range correlations of chain motion due to the short length of statistical segments.
a)
22
b)
22 20
20
TEM SAXS
drigid [nm]
Lp [nm]
temperature [ C]
c)
1.6 1.4 non-annealed annealed
d)
1.2 non-annealed annealed 1.0
dsemi-rigid [nm]
dsoft [nm]
0.8
temperature [ C]
Fig. 4.7. The temperature dependence of the long period from NMR (a), the thickness of rigid (b), semi-rigid (c), and soft (d) domains of HDPE. The experiments were performed on low-field NMR spectrometer for a compression moulded plate before and after annealing at 121 C for 100 hours. The shaded area shows the lamellar thickness measured by SAXS and TEM.
38
4.8 The effect of annealing on the thickness of domains and molecular mobility
In order to determine the effect of annealing on the thickness of the rigid, semi-rigid, and soft domains, a compression moulded HDPE plate was annealed at 121 C for 100 hours. The thickness of domains was determined at 70 C, 100 C and 120 C and the results for the annealed sample are compared to those of the sample not annealed (Fig. 4.7). Annealing causes a significant decrease in the thickness of the interfacial layer and the soft amorphous phase, and an increase in the lamellar thickness, which can clearly be seen for the results measured at 120 C. The annealing causes only minor changes in molecular mobility (Fig. 4.5b). The annealing study suggests the following mechanism of morphological changes due to annealing at elevated temperatures. Upon approaching the melting temperature, the molecular mobility increases both in the amorphous regions, the crystal-amorphous interface, and in the crystalline phase (-relaxation) (Fig. 4.4). Previous
13
suggested two types of crystalline environments with distinctly different molecular mobilities: a more perfect one and one with defects [Hil1998, Hu2000]. It might be suggested that chain mobility increases at elevated temperatures in the less ordered, surface layer of lamellae, as follows from a decrease in the lamella thickness above 100 C for the non-annealed sample (Fig. 4.7b). An increase in the molecular mobility is accompanied by partial melting of small crystals and less ordered fragments of the lamellae. An increased mobility in the amorphous phase and chain diffusion in- and out of the crystals ( - relaxation) [Sch1991] facilitate to some extent morphological changes towards a more thermodynamically stable, better-ordered crystalline structure and thicker lamellae. It should be mentioned in this respect that the rootmean-square displacement of chains in PE lamellae amounts to more than 10 nm at 100o C within 100 seconds [Sch1991]. Chain rearrangements upon prolonged annealing at 121 C for 100 hours result in a decrease in molecular mobility in the crystalline phase, as follows from a slight decrease of T2rigid during annealing and large increase in the melting enthalpy (from 232.8 J/g to 273.9 J/g) at only 1 wt% increase in the crystallinity (Fig. 4.5b). This might suggest perfection of the crystalline order. Lamellae thickening and a slight increase in the crystallinity would cause additional slippage of chain entanglements towards the inner part of inter-lamellar amorphous regions causing additional immobilization of the soft fraction of the 39
amorphous phases, as can be concluded from the decrease of the thickness of this layer (Fig. 4.7c) and the decrease in T2l (Fig. 4.5b). Thus, our observations reveal that annealing is accompanied by a continuous shift of the crystal-amorphous interface towards the inner part of the amorphous regions and a reduction of the thickness of the amorphous layer [Che1994, Mat2001, Bas1973, Str1997 ].
4.9 Conclusions
The effects of temperature and annealing on the following molecular and morphological characteristics of HDPE were studied by time- and frequency- domain proton solid-state NMR: (i) the amount of rigid, semi-rigid, and soft fractions of HDPE, (ii) the thickness of the domains with different molecular mobilities, and (iii) the molecular mobility in these domains. This study focused on two areas, namely (1) the improvement of low-field solid-state NMR methods for determining the thickness of domains in heterogeneous polymers, and (2) a better understanding of the phenomena that occur upon annealing of HDPE. (1) The novel spin-diffusion experiments that are based on the double-quantum dipolar filter provided reliable information about the thickness of lamellae, crystal-amorphous interface and soft fraction of the amorphous phase. The applied method allows a better discrimination between the HDPE phases and, therefore, a more accurate analysis of the thickness of the domains. In order to improve the accuracy of the method, the temperature dependence of the spin diffusivities in all three phases was determined, which allowed us to obtain reliable information about changes in the domain thickness as a function of temperature. The lamellae thickness and the long period determined by NMR, are in good agreement with the results of classical methods, i.e. TEM and SAXS, which proves the reliability of the NMR method. This study has also showed that spin-diffusion experiments with advanced dipolar filters can be performed with a low cost, low-field NMR spectrometer, which broadens the range of applications of this equipment. (2) This study shows that the three-phase model is the most appropriate description of the phase composition in HDPE at least as revealed by 1H NMR data. The spin-diffusion experiments show that rigid and soft fractions of HDPE are separated by a thin layer with intermediate molecular mobility. The thickness of this interface is nearly constant in the temperature range from 70 C to 120 C and its value is comparable to the length of the statistical segment of the PE chains. This suggests that the interfacial layer largely originates 40
from constraints on rotational and translational motions of chain fragments adjacent to the lamellae surface. The crystalline phase, the semi-rigid crystalline-amorphous interface, and the soft fraction of the amorphous phase exhibit a larger contrast in molecular mobility at temperatures above 90 C. Upon increasing the temperature, the molecular mobility begins to increase in the inner/softer part of the amorphous phase towards the lamellae surface. Annealing causes a significant decrease in the thickness of the interfacial layer and the soft amorphous phase, and an increase in the lamellar thickness. It is suggested that annealing is accompanied by structural reorganizations at the frontier between the crystalline and amorphous phases, and these reorganizations are largely influenced by the chain dynamics in the crystalline and amorphous phases. Annealing leads to improving perfections of the crystalline order (presumably in a layer close to the lamellae surface), and a continuous shift of the interface towards the inner part of the amorphous regions reducing the thickness of the soft amorphous layer. These observations are in agreement with a previously proposed mechanism of partial melting and surface crystallization upon annealing of HDPE at premelting temperatures [Str1997].
41
42
Chapter 5
Isotactic Polypropylene (iPP)
5.1 Introduction and motivation
The physical and mechanical properties of semi-crystalline polymers (like isotactic polypropylene), and their practical lifetime, are significantly influenced by both the changes in the degree of crystallinity [Pop1987, Hal1972, Boy1979, Boy1983,Boy1984, Boy1985, Str1980, Man1964, Man1985, Neg1992, Wor1996], and the orientation distribution of the crystalline volume portion [Bel1993, Gos1996]. Both quantities typically change when the sample is exposed to temperatures well above the glass transition due to annealing [Hik1997, Alo1999, Aki1999, Mar2000, Iij2000, Mai2000, Weg2000, Ali2001, Lab2001]. This means that the degree of the crystallinity and the orientation distribution depend on processing conditions, crystallization, annealing, and thermal history in such materials. It is known that annealing of semi-crystalline polymers after fabrication causes changes in the structure and properties of the materials. This process has been employed to improve the final properties of polymers via healing of defects and diminishing residual stress and strain. In addition, lamellae thickening and rearrangement of the chain during heating also occur upon annealing. For establishing structure-property relationships of semi-crystalline polymers, the characterization of lamellar thickness in bulk materials is a necessity, because their thermal and physical properties depend largely on the characteristic of the lamellae. Traditionally, transmission electron micrograph (TEM) and small-angle X-ray scattering (SAXS) are two major techniques used to determine the lamellar thickness. The former offers the advantage of direct access to the morphology. SAXS is a well-developed method to quantitatively determine the thicknesses of alternating layers of the crystalline and amorphous regions that are formed in semi-crystalline polymers. The morphological parameters can be deduced from SAXS data based on a model calculation. Although an indirect approach is applied, SAXS provides volume average characteristic. Proton NMR spin-diffusion experiments have been used to provide information about the thickness of domains with different molecular mobility in polymers [Bri1993, Pac1984] [Hu2000, Bud2003, Bud2004, Van1996]. The main aim of this chapter is to study the change in the rigid fraction/crystallinity, semi-rigid, and soft fractions, and lamellae thicknesses for both annealed and non-annealed 43
iPP samples. This was done using a combination of DSC, SAXS, WAXD, TEM, and 1H solid-state NMR, including spin-diffusion NMR. The 1H NMR spin-diffusion methods, which explore both double-quantum (DQ) dipolar filter [Bud2003, Ba1998, Che2005] and the Goldman-Shen (GS) filter [Gol1961, Dem1996], were used for the first time to provide a more accurate analysis of the domain thickness of iPP. It is thought that TEM, SAXS, and NMR are complementary techniques, providing morphological evidence on the nanometer scale. The most probable morphology was established by comparing the 1H spin-diffusion data with the multidimensional evaluation of the magnetization by spin-diffusion. The results are compared under the given conditions, which include the annealing time and annealing temperature. A simple model of crystal thickening as a function of the annealing time is discussed. The correlation between transverse relaxation rates and thickness of the rigid domain is presented, i.e., microscopic and mesoscopic properties.
5.2 Materials
Isotactic polypropylene homopolymer was obtained from SABIC Europe BV and used in our experiments. The material was injected in an Engel 45A machine. The melt temperature was set to 235 C, the holding pressure to 40 MPa, the holding time to 20 s, the cooling time to 20 s, and the overall cycle time to 49.5 s. For the annealing study, the sample was further annealed in the NMR probe at 134 C, 143 C and 153 C, for several times between 15 minutes and 30 hours.
Litvinov and Soliman showed that annealing of iPP gives rise to shoulders in the DSC thermograms [Lit2005]. In order to compare the annealed samples with non-annealed iPP, an offset temperature is defined in this study. The offset temperature corresponds to a drastic increase of the heat capacity due to melting of the polymer (see Fig. 5.1b). Offset temperatures of 144 C and 153 C are observed for iPP annealed at 143 C and 153 C for 10 hours, respectively. An apparent offset temperature of 134 C (increase of Cp in comparison to the baseline) is observed for non-annealed iPP. The determined offset temperatures for the annealed samples correlate with the applied annealing temperatures. These observations are in agreement with the NMR data discussed below.
a) heat capacity, cp [kJ.kg .K ]) 4 100 crystallinity, W [%] 80 3 60 40 20 1 0 50 100 150 temperature [C] 0 200
-1
cp non-annealed iPP
o cp iPPannealed at 143 C o cp iPPannealed at 153 C
-1
cp,a cp,c
3
134 C 144 C 153 C
2 120
c) heat capacity, cp [kJ.kg .K ] 4 2 0
160
-1
-1
-2 -4
temperature [C]
45
d)
heat capacity, cp [kJ.kg .K ]
-1
12
-1
0 40
80
120
160
200
temperature [C]
Figure 5.1. (a) Heat capacity ( c p ) and crystallinity (W) data recorded during the first heating of non-annealed iPP (black solid and dashed lines), annealed at 143 C iPP (red solid and dashed lines), and annealed at 153 C iPP (blue solid and dashed lines). The annealing time was 10 hours. The green and orange dotted lines represent the heat capacity for the crystalline (cp,c) and amorphous (cp,a) phases, respectively. (b) Heat capacity and onset of melting for non-annealed iPP (black line), annealed at 143 C (red line) and annealed at 153 C iPP (blue line), respectively. The annealing time was 10 hours. (c) Heat capacity (solid lines) and crystallinity (dashed lines) recorded during the second heating (cp,h,a and cp,h,c, Wc,h) and cooling (Wc,c, cp,c,a and cp,c,c) of iPP. The dotted lines represent the heat capacity for the amorphous and crystalline phases for the second heating and cooling, respectively. (d) Heat capacity curves recorded for the second heating curves for non-annealed iPP (black line), annealed at 143 C (red line) and annealed at 153 C iPP (blue line). The annealing time was 10 hours.
A complete heat capacity analysis is performed. The extreme states in which the material can be found (100% crystalline and 100% amorphous) are referred to as reference states. These data are reported for a large number of polymers in the ATHAS Data Bank [Pyd1994]. Mathot has shown that the polymer can be regarded as consisting of regions having the same thermal properties as the reference states, and a simple two-phase model can be used in order to determine the enthalpy-based mass fraction crystallinity [Mat1944]. This means that no transition layers between the regions and no contributions from interfaces are taken into account for the analysis of DSC data in this study. We observed that the crystallinity of iPP, determined at 80 C (W80C), increases as expected with the annealing temperature: W80C = 50 wt% for non-annealed iPP, 59 wt% for iPP annealed at 143 C and 61 wt% for iPP annealed at 153 C, (Fig.5.2).
46
Wc non-annealed iPP 1st heating Wc annealed iPPat 143oC Wc annealed iPPat 153oC Wc non-annealed iPP 2nd heating
temperature [C]
Figure 5.2. Crystallinity of iPP at the first heating curves (solid lines) of non-annealed iPP (black), iPP annealed at 143 C (red) and 153 C (blue). The annealing time was 10 hours. The crystallinity at second heating curve is shown by the orange dashed line.
After erasing the thermal history by melting, the crystallinity of iPP (Fig. 5.2 - second heating curve; the dashed, orange line) is somewhat higher than the crystallinity of nonannealed iPP (Fig. 5.2 - first heating curve; the straight, black line), but still lower than the crystallinity of both annealed iPP samples (W80C = 55 wt% for the second heating curve). Thus, the crystallinity of iPP derived from the first heating curve shows a thermal history different from the one derived from the second heating curve (Figure 5.1d). The end of the melting behaviour of the annealed samples is located at a higher temperature than that of the second heating curve (indicated in Fig. 5.2 by the green arrow). This is expected when crystal thickening and crystal perfection occur. NMR experiments also show an improvement of crystal perfection upon annealing, as will be shown below. Crystal thickening is observed by NMR and TEM during annealing of iPP samples. Thus, the results of NMR and TEM are in line with the heat capacity data.
47
b)
100 nm
100 nm
c)
100 nm
Figure 5.3. TEM image for non-annealed iPP (a), iPP annealed at 143 C (b), and 153 C (c). The annealing time was 10 hours. The scales of the images are show in the figures.
500
48
b)
1500 non-annealed
1000 I(2)
500
0 5 10 15 2 20 25 30
Figure 5.4. a) One-dimensional WAXD intensities obtained via integration of two-dimensional WAXD patterns for non-annealed iPP, and iPP annealed at 143 C and 153 C for 10 hours. b) Deconvolution of WAXD pattern for crystalline and amorphous phases for non-annealed iPP.
Integration of the 2D-WAXD pattern over a pie-shaped area yielded the intensity profiles displayed in Fig. 5.4a. In all cases, the sequence of the detected diffraction maxima is consistent with the monoclinic -structure of polypropylene. The contributions arising from the crystalline and the non-crystalline components are extracted via peak fitting of the diffraction pattern (Fig.5.4b). The latter yielded the (X-ray) crystallinity which showed a 9% increase from a value of 0.49 0.05 for the non-annealed system to 0.58 0.05 for a sample annealed at 153 C. These values are in good agreement with DSC crystallinity.
15 12
2
I(q)*q
Figure 5.5. One-dimensional SAXS intensity I(q)=I(q)*q2 recorded for non-annealed iPP and iPP annealed for 10 hours at 143 C.
The SAXS patterns reveal well-defined interference maxima owing to a quasi-periodic arrangement of the crystalline domains. More details about the scattering and correlation functions are discussed in Ref. [Hed2007]. The one-dimensional scattering intensities 49
I (q ) * q 2 are shown in Fig. 5.5, where q = (2 / ) * sin , is the wavelength of the X-ray
radiation and, is the scattering angle and q is the scattering vector. It follows from the results in Fig. 5.5 that annealing has a strong influence on the lamellae thickness. After annealing, the position of the interference maximum shifts towards lower scattering angles which indicates an increase of the effective long period. This is also reflected by the linear correlation functions, calculated from the scattering intensities (Fig. 5.6).
non-annealed o annealed at 143 C
1.0
g(r)
0.5
0.0
-0.5 0
10
20 r [nm]
30
40
Figure 5.6. Normalized linear correlation function (r) revealing an increase in the long period and phase dimensions due to annealing at 134 C, 143 C and 153 C for 10 hours.
In this case, the position of the first side maximum indicates the changes in the value of the long period following different thermal histories. Moreover, the inner part of the correlation function provides insight into the thickness of one of the phases. However, it cannot be concluded directly whether or not the corresponding value reflects the thicknesses of the crystalline or the amorphous domains.
50
5.6 Solid-state
NMR
study
of
phase
composition,
molecular mobility, and domain thickness 5.6.1 Temperature dependence of phase composition and chain mobility
A quantitative analysis of the FID and wide-line NMR spectra in terms of crystalline phase, semi-rigid phase, and soft phase was performed at a temperature well above Tg. The amounts of the rigid, semi-rigid and soft fractions of iPP obtained from low- and high- field NMR are shown (Fig. 5.7). At room temperature, a significant fraction of the soft phase is rigid, and contributes to both the intensity of the broad line of the NMR spectra, and the short T2 relaxation component of the FID. At elevated temperature, the amount of the rigid fraction, which comprises the crystalline phase and the rigid fraction of the soft phase, slowly decreases. While the amount of the soft fraction increases under the same conditions, the amount of the semi-rigid fraction is almost constant in the temperature range from 70 C to 100 C (Fig. 5.7).
a) 80 60 content [%] 40 20 0
rigid fraction semi-rigid fraction soft fraction
b) 80 60 content [%] 40 20 0
rigid fraction semi-rigid fraction soft fraction
40
60
80
40
60
80
temperature [ C]
temperature [ C]
Figure 5.7. Temperature dependences of the amounts of rigid, semi-rigid, and soft fractions of non-annealed iPP measured with low-field (a) and high-field (b) NMR spectrometers.
The data in Fig. 5.8 show nearly constant values for T2rigid and 1/2rigid with temperature increasing up to 100 C, indicating that the molecular motion in the rigid fraction is almost unchanged. The small increase in T2rigid above 100 C may be related to the motion in the crystalline phase ( relaxation process) which occurs prior to melting and to the thermal 51
expansion of crystal lattice [Sch1994]. In the temperature range from 40 C to 60 C, the transverse relaxation time of the semi-rigid and soft fractions increases largely, whilst semirigid 1/2,
and soft1/2 largely decrease (Fig. 5.8). These changes are due to the high-frequency
(ca. 10 kHz) manifestation of the glass transition of the amorphous phase. The temperature dependences of the T2soft and of 1/2soft can be correlated by the relationship 1 / 2
soft
1 , T2
valid for a Lorentzian line shape (Fig. 5.8). In the temperature range from 100 C to 140 C, an increase in the T2 relaxation time of the semi-rigid and soft fractions is due to partial melting of isotactic polypropylene. That reduces hindrances of the chain mobility in the soft phase.
a) 0.08 0.06 1/T2 [s ]
-1
b) 80 60
1/2 [10 Hz] 1/2 rigid 1/2 semi-rigid 1/2 soft
40 20 0 20
40
40
Figure 5.8. Temperature dependence of the 1H 1/T2 relaxation rates (a), and line-width 1/2 (b) for nonannealed iPP. The assignment of the parameters to the rigid, semi-rigid, and soft fractions of iPP is shown in the figure.
In conclusion, the temperature range from 70o C to 100 C appears to be most suitable for accurate determination of the phase composition. In this temperature range, the differences in molecular mobility in different fractions of iPP are higher, and annealing can be lately avoided during the NMR experiment, as shown below. In the temperature range from 70o C to 100 C, the amount of the rigid fraction is largely independent of temperature, and its values is closed to that of the crystallinity measured by DSC and X-ray scattering, as discussed below.
52
5.6.2 The effect of annealing temperature and annealing time on the phase composition and molecular mobility
To determine the temperature at which annealing causes changes in the morphology, two different types of annealing studies were performed. The annealing was studied at isothermal conditions in the temperature range from 70 C to 130 C in real time. The data were recorded each hour for 30 hours. No annealing effects were detected by NMR for iPP samples exposed to 70 C, 90 C and 110 C for 30 hours. In the temperature range from 110 C to 130 C, small changes in the molecular mobility and the phase composition were observed over a time of 500 hours. Annealing at 130 C for 500 hours causes 5 wt% increase in the amount of the rigid fraction at the expense of the soft phase. The amount of the semi-rigid phase remains almost constant (Fig. 5.9a). Figure 5.9b show that the molecular mobility in the crystalline phase of iPP is hardly affected by annealing at 130o C, whereas the molecular mobility of the semi-rigid fraction decreases and the molecular mobility in the soft fraction increases during the annealing.
a) 62 60 content [%] 58 250 rigid fraction semi-rigid fraction soft fraction T2 [s] b)
25 20 15
100
200
300
400
500
Figure 5.9. (a) The amounts of rigid phase, semi-rigid interface, and soft fraction of iPP as a function of the annealing time at 130 C. (b) The effect of the annealing on molecular mobility, as determined by the T2 relaxation time for the different fractions of iPP. Larger amplitude and/or frequency of molecular motions lead to longer T2 values. The relaxation characteristics were determined by the analysis of the FID (rigid and semi-rigid fractions), and Hahn-echo (soft fraction), as described in experimental section. The experiments were performed with low-field NMR spectrometer.
From the above results, we can conclude that exposure of iPP to temperatures between 100 130 C influences the NMR results, due to annealing of the samples during NMR experiments. These experiments were used to determine the maximum temperature at which no annealing occurs. 53
The effects of the annealing time and the annealing temperature on the phase composition and molecular mobility were studied for iPP samples annealed at 134 C, 143 C, and 153 C. The annealing times are taken in the range from 15 minutes to 30 hours. After annealing at the indicated temperatures, the NMR experiments were performed at 70 C. A fast increase in the amount of the rigid fraction is observed during the first five hours of annealing at 134 C, 143 C, and 153 C (Fig. 5.10a). The amount of the rigid fraction increases by a few wt% at the expense of the soft fraction. The amount of the semi-rigid fractions remains almost constant. It is also observed that the molecular mobility in the rigid and semi-rigid fractions of iPP decreases, while the molecular mobility of the soft fraction increases during the annealing. This increase is due to the better organised morphology of iPP during annealing. This is shown by the decrease in the T2 values for the rigid fraction (T2rigid) (Fig. 5.10b), the semi-rigid fraction (T2semi-rigid), and the increase in the T2 values for soft fraction (T2soft). The observed changes during annealing can take place for the following reasons: formation of thin lamellae between the primary lamellae, lamella thickening, and improvement of the ordering in different phases [Weg2000, Ito1992].
b)
a)
66
14.0
rigid fraction [%]
[s]
T2
rigid
13.6
13.2
Figure 5.10. The effect of annealing time and annealing temperature on (a) the amount of the rigid fraction (% T2rigid), and (b) 1H relaxation time (T2rigid) of this fraction of iPP.
54
5.6.3
Comparison
of
the
amounts
of
rigid
SAXS (vol%)
crystalline amorphous
DSC (wt%)
crystalline amorphous
56 67 70
36 26 24
8 7 6
51 55 58
49 45 42
51 59 61
49 41 39
The reasons for the apparent differences between the results obtained by NMR, DSC, and SAXS have been discussed in the previous chapter [Hed2007].
5.6.4 Comparison of the domain thicknesses measured by the DQ and the Goldman-Shen dipolar filters in spindiffusion experiments
The confidence in both the values of the domain thickness and the effective dimensionality of the 1H spin-diffusion process can be enhanced if the experiments are performed on the same iPP sample using different dipolar filters. In this case, we employed DQ and Goldman-Shen (GS) filters. These dipolar filters select the initial magnetization in the rigid and amorphous domains, respectively.
55
5 (td)
10
1/2 1/2
15
20
4 (td)
1/2
[ms ]
8 12 1/2 [ms ]
16
Figure 5.11. Proton spin-diffusion decay and build-up curves for the rigid () and amorphous () fractions of non-annealed iPP recorded using the DQ (a) and GS (b) dipolar filters. These filters select the magnetization of the rigid and amorphous fractions of iPP, respectively. The time-domain data were recorded at 70 C with a lowfield NMR spectrometer and fitted with one Abragam function and one exponential function. The filter time for the GS was set to 100 s. The solid lines represent the best fit for the 1D solutions of the spin-diffusion data for the two-phase morphology.
The measurements of 1H spin diffusion using both dipolar filters were performed on the non-annealed iPP sample at 70 C with low-field NMR. The spin-diffusion build-up and decay spin-diffusion curves are shown in Fig.5.11. From these data it is evident that the spin-diffusion process reaches the quasi-equilibrium state after a diffusion time of the order of td 70ms for both filters. Moreover, for the DQ filter, the quasi-equilibrium of the relative magnetization for the rigid and amorphous phases equals 0.73 and 0.27, respectively. In the case of GS dipolar filter, the corresponding phase fractions are 0.79 and 0.21. In the limit of experimental errors, these two results prove that both dipolar filters provide similar values of domain thickness. Small differences exist because the efficiency of the dipolar filters in selecting the magnetization just from one domain is not 100%, and the shape of the relaxation components biases the accuracy of the FID deconvolution procedure. The thicknesses of the rigid (dr) and amorphous (da) domains are determined from fits of the spin-diffusion data (Fig. 5.11) with the 1D solutions of the spin-diffusion equations. The approximation of two phases should not have a large effect on the extracted domain thicknesses, taking into account the small amount 4 wt% of the soft fraction. The time dependence of the analytical spin-diffusion signals for the GS filter can be obtained from those of the DQ filter, by interchanging the indices r and a of the quantities present in the spin-diffusion solutions. The domain thicknesses obtained by fits of the spin-diffusion data of Fig. 5.11 are given in Table 5.2.
56
Chapter 5. Isotactic Polypropylene Table 5.2. The thicknesses of the rigid (dr) and amorphous (da) domains of non-Annealed iPP measured at 70 C using spin-diffusion experiments with double-quantum (DQ) and Goldman-Shen (GS) dipolar filters (Fig. 5.11).
dipolar filter DQ GS
dr [nm] 10 9
da [nm] 2.5 3
The time domain data were fitted with a combination of one Gaussian and one exponential function, and with a combination of one Abragam and one exponential function. The domain thickness with values closest to that measured by SAXS, is that obtained from the free induction decay deconvoluted with the sum of one Abragam and one exponential function. Moreover, both dipolar filters provide a volume ratio of the rigid and amorphous domains that is close to the mass fraction of these phases, as determined from the FID at different temperatures, i.e., d r / d a 2 5 and %T2rigid / %T2amorphous 3 4.8 . These results provide additional proof of validity of the method for domain thickness determination, as well as the type of morphology. The thicknesses of the iPP domains determined by time-domain and frequency domain spin-diffusion experiments are in good agreement (Table 5.2).
57
5.6.6 The effect of annealing temperature and annealing time on the domain thickness of iPP samples
The long period Lp,, and the thickness of the rigid dr and amorphous da domains, was studied as a function of annealing temperature and annealing time. The spin-diffusion experiments were performed at 70 C for the iPP samples, annealed at three different temperatures, and different annealing times. The values of the long period before and after annealing, determined by NMR spin-diffusion experiment and SAXS, are in a good agreement (Table 5.4).
Table 5.4. The long perioda of non-annealed iPP and iPP that was annealed for 10 hours at 143o C and 153o C as determined by NMR and SAXS. non-annealed annealed at 143o C annealed at 153o C NMR (70 oC) 12.5 nm 19.9 nm 21.3 nm SAXS data 12.9 nm 21 nm 21.8 nm SAXS, linear 12.5 nm 20 nm 21 nm correlation function a The average errors are of the order of 15%
Figure 5.12 shows the changes of Lp, dc, and da with annealing time. A rapid initial increase in the long period and in the thickness of the rigid domains is observed for the iPP samples during first 10 hours of annealing (Fig.5.12ab). The rapid increase in the crystal thickness with annealing is confirmed by the increase in the crystallinity (Fig. 5.10a, Table 5.1, and Table 5.4). For long annealing times, ranging between 10 30 hours, Lp and dc do not seem to change significantly, showing a stabilization of the iPP morphology (Fig.5.12ab) [Pou1998].
a) 22 20
Lp [nm]
b) 20
18 16 0
dr [nm]
16
14
58
2.5 da [nm]
2.0
Figure 5.12. Changes of the (a) long period Lp, (b) thickness dr of rigid domains, and (c) thickness da of amorphous domain of iPP as a function of annealing temperature and annealing time. The domain thicknesses were determined from 1H spin diffusion experiments exploring a DQ filter. The experiments were performed with a low-field NMR spectrometer.
The thickness of the amorphous domains slightly decreases with increasing the annealing temperature and annealing time (Fig. 5.12c). Chain dynamics and thermodynamic factors play an important role in the annealing process. The growth of lamellae involves translational motion of the polymer chains [Fer1984]. Faster chain diffusion through iPP crystals, and higher chain mobility in the amorphous phase upon increasing the temperature facilitate the structural reorganization towards thermodynamically more stable thicker crystals [Sch1991, San1973] Previous studies have also shown that the annealing of iPP samples at sufficiently high temperature increases the thickness of the lamellae [Pou1998, Mar1998, Fer1984].
(5.1)
where y = dr/dr0, dr0 is the equilibrium thickness of the crystal, t = kd r20 and k is a proportionality constant. The temperature dependence of the thickening rate should enter the theory through k, or equivalently t . When t is assumed to be independent of dr, the above equation can be integrated analytically, and the transverse crystal thickness dr exhibits a sigmoidal shape as a function of log(t a / t ) [San1973, Sat1961, Fis1962, San1974, Pet1963, Pet1967]. At the intermediate values of log(t a / t ) , the crystal thickness increases approximately linearly with a higher rate at higher temperatures. Therefore, in the intermediate range of the annealing time t a , the rigid domain thickness d r measured by spindiffusion can be described by the relationship d r d r 0 + B log(t a / t ) . (5.2)
For an annealing time t a = t we get from the above equation d r d r 0 . When t a = e t we get the significance of the constant B, i.e., B d r d r 0 . The dependence of the rigid domain thickness for annealed iPP at three different temperatures is shown in Fig. 5.13 as a function of log(t a ) . At all annealing temperatures, the data can be fitted with a straight line in agreement with Eq. (5.2). Therefore, the process of crystal thickening upon annealing for iPP can be described in the linear approximation by the Eq. (5.2) with can be derived from Eq 5.1 [San1973, San1974]. A strong temperature dependence of the thickening rate is not detected (cf. Fig. 5.13) probably due to the limited range of annealing temperatures. Moreover, the thickening rate is primarily a function of
0 0 undercooling T = Tm T , where Tm is the equilibrium melting point of the crystal (about
60
20
dr [nm]
18
16
Figure 5.13. Change of the thickness d r of the rigid domains with the annealing time t a at three different annealing temperatures. The straight lines are a least-squares fit of the data using Eq. (5.2).
5.6.8 Correlation between the 1H transverse relaxation rate and the domain thickness of the crystalline domains.
The microscopic properties of iPP are represented by several parameters that describe the polymer chain dynamics. These dynamics correlate with the local structure on the nanoscale range and are revealed inter alia by the 1H transverse magnetization relaxation rates. The morphology and domain thickness characterize the materials on the mesoscopic scale, i.e., on the scale of tenth or hundreds of nanometers. The values of the 1H effective transverse magnetization relaxation rate ( 1 / T2rigid ) for the rigid domain is related to the residual second van Vleck moment M 2 , i.e.,
1 / T2rigid M 2 . This is evident from the fact that the free induction decay G (t ) of the rigid
domains in the short time domain is described in a good approximation by the Gauss function
(5.3)
Consequently, faster transverse relaxation rates correspond to larger values of M 2 due to larger strength of the dipolar couplings. This is the case of slower chain motions, and 61
increased interchain dipolar interactions due to denser chain packing and/or better crystal organization.
1 hour 5 hours 10 hours 20 hours
[s ]
-1
153 C
0.076
143 C
0
1/T2
rigid
0.074
0.072
134 C
0
14
16
18 dr [nm]
20
Figure 5.14. Correlation between the 1H transverse magnetization relaxation rate ( 1 / T2 ) and the thickness of the rigid domains for iPP that was annealed at 134 C, 143 C, and 153 C for 1, 5, 10, and 20 hours. The NMR measurements were performed at 70 C using a low-field NMR spectrometer. The lines are a guide for the eye.
rigid
A phenomenological attempt to correlate 1 / T2rigid and d r for different annealing temperatures and annealing times is shown in Fig. 5.14. The transverse magnetization relaxation and spin-diffusion measurements were made at the same temperature of 70 C. Therefore, the changes in the values of 1 / T2rigid are due to the crystal thickening and perfection of the crystalline order during annealing. The theory of polymer annealing [San1973, San1974, Pet1963, Pet1967] predicts that for a given annealing temperature the crystalline domain thickness increases linearly with log(ta) (see Eq. (5.2)), where ta is the annealing time. This is evident from the values of d r measured as a function of ta at different annealing temperatures (Fig. 5.14). Only small changes in the values of 1 / T2rigid are detected at each annealing temperature as a function of annealing time. This parameter increases slightly with annealing time due to an increase in the density (perfection) of the crystalline regions during annealing [San1973]. This leads to an increase in 1 / T2rigid , due to an increase in the 1H dipolar couplings, and consequently in M 2 . Higher annealing temperatures lead to more perfect and better-packed crystals in which fast small-angle fluctuations are reduced. Moreover, we note that this major effect occurs rapidly, and only minor further changes are observed during the course of annealing. This shows that there is no simple correlation between 1 / T2rigid and crystalline the domain thickness.
62
5.7 Conclusions
Changes in phase composition (namely the amounts of rigid fraction/crystallinity, semi-rigid, and soft fractions), molecular mobility and domain thickness of both annealed and nonannealed iPP samples were investigated by 1H NMR, SAXS, and DSC at different temperatures and different annealing times. The amounts of the rigid fraction as determined by NMR, exceeds by 5-10% the crystallinity from DSC and SAXS. This difference can quantitatively be explained by the effect of the interface, which is not taken into account in the analysis of the DSC and SAXS data. According to TEM, the morphology of non-annealed iPP is mainly lamellar. For iPP samples that have been annealed at different temperatures and different times, the crosshatching morphology can be observed. However, the predominant morphology remains lamellar. This is confirmed by the comparison of the phase composition with the domain thickness ratio that is determined by simulation of the results of the spin-diffusion data using 1D, 2D, and 3D solutions of the spin-diffusion equations. In an approximation of the twophase model, the long periods obtained by NMR and SAXS are in good agreement with each other. An important feature of the spin-diffusion experiments reported in this investigation is the temperature dependence of the proton spin diffusivities, which decrease with increasing temperature. The temperature dependence of the spin diffusivity of the amorphous phase was derived from the spectral line width. In the approximation of a single correlation time described by the Williams-Landel-Ferry (WLF) equation, the temperature dependence of the spin diffusivities was established, and the obtained WLF coefficients are in the range that is typical for the amorphous phase in semi-crystalline polymers and amorphous polymers. The thicknesses of the rigid domains for iPP that has been annealed at different temperatures and different times has been discussed as a function of the logarithm of the annealing time. A linear relationship holds for all cases in good agreement between theory and the experimental results for other polymers [San1973, San1974]. The existence of a phenomenological correlation between microscopic and mesoscopic properties of iPP is reported. Microscopic properties are described by 1H transverse relaxation rates or second van Vleck moment characterizing mainly the proton dipolar network. The morphology and domain thickness characterize the polymer on the mesoscopic scale. The correlation is complex and explained by the improvement of the crystalline order during the crystal thickening.
63
64
Chapter 6
Aging on Isotactic Polypropylene
6.1 Introduction and motivation
Physical aging of semi-crystalline polymers, which has been studied extensively in the past [Cru1984, Cha1980, Tom1996, Uzo1997, Ibh1996, Hel1997] from many aspects, refers to the evolution towards an equilibrium state and the associated changes in the morphological organization and physical properties. Above the glass transition the physical aging at different temperatures and different times involves the densification of the amorphous fraction (reduction in free volume) [Son1987], constraints in the chain mobility [Aga1981], lamellar perfection or densification, lamellar thickening, crystallization of constrained polymer chains forming small crystals or fringed micelles, and melting-recrystallization. The exact origin of the changes in iPP samples (and in other semi-crystalline polymers) with aging time and aging temperature is still under dispute [Cru1984, Cha1980, Tom1996, Uzo1997, Ibh1996, Hel1997, Son1987, Aga1981]. One reason of the dispute lies in the limitations of different techniques used to probe the structure of these materials and the discrepancies between their results. Differential scanning calorimetry (DSC) rarely provides similar values for crystallinity as other techniques [Isa1999, Zho1997]; wide-angle X-ray diffraction (WAXD) requires an estimation of the amorphous fraction and the lattice imperfection factor [Rul1961]; and smallangle X-Ray diffraction (SAXS) is extremely model dependent [Bal1989]. It should be noted that the problems with SAXS analyses are enhanced in the case of iPP due to the crosshatched lamellar morphology [Oll1989]. One feature that is generally agreed upon by most researchers as important for aging behaviour is whether the annealing temperatures are above or below the crystalline dielectric relaxation c. The crystalline relaxation c is associated with the rotational-translational motion travelling along the crystal stem [Boy1985, Boy1985, Spi1991, Hu1999] that can grow from the soft fraction in a crystal thickening process. The responsible mechanism for crystallinity increase due to aging is the secondary crystallization by which the new lamellae are inserted between those formed during the primary crystallization. This mechanism occurs since the regions remaining amorphous still have the potential to partially crystallize. The 65
newly formed lamellae are constrained by those formed during primary crystallization. They are thicker as they are formed at higher temperatures. Because an additional driving force is needed, the newly formed lamellae at lower temperature upon further non-isothermal cooling are thinner. They may thicken if a low temperature active process is plausible. It is not only a matter of aging temperature, with respect to the glass transition temperatures of the crystalline and amorphous phases, for the new formed lamella, but also depends on the initial iPP morphology. The first aim of this chapter is to study in detail the effects of aging temperature and aging time on the phase composition, crystallinity, melting and crystallization temperatures, chain mobility and mechanical properties of injection molded, crystallized from the melt, and quenched homopolymeric iPP by 1H low field solid state NMR, DSC and X-ray. The second goal of this chapter is to show that the 1H low-field solid-state NMR technique is capable of detecting the small changes in the phase composition and chain mobility due to aging at different times and different temperatures. Moreover, a correlation between the changes in the NMR, and DSC parameters due to aging and mechanical data is presented and discussed.
6.2 Materials
Three homopolymeric isotactic polypropylene (iPP) samples were used in our investigations. All the samples are commercial Ziegler-Natta-catalyzed PP manufactured by SABIC Europe BV. The first sample is an injection molded homopolymeric iPP. The materials were injected in an Engel 45A machine. The melt temperature was set to 235 C, the holding pressure to 40 MPa, the holding time to 20 s, the cooling time to 20 s, and the overall cycle time to 49.5 s. The second sample is melt-crystallized homopolymer iPP. The injection molded plates were melted at 230 C for 30 minutes in an oven under nitrogen flow, in order to avoid the oxidation of the samples, and finally crystallized at a slow cooling rate. The third sample is a quenched homopolymeric iPP sample. The sample has been brought to 250 C, kept 30 minutes at that temperature and than quenched from the melt at -195 C between two metal plates. For the aging study at 28 C, all the homopolymeric iPP samples were further annealed in the NMR probe, for times between 10 minutes and 1000 hours after preparation. For the aging study at different temperatures the iPP samples crystallized from the melt was annealed in the NMR probe for the temperature range from 70 C to 130 C, for times between 10 minutes and 150 hours after preparation. The temperature of the NMR experiments coincides with the annealing temperature. All NMR measurements were repeated 66
with an interval of one hour. The samples were not removed from the NMR probe for 150 hours and 1000 hours, respectively.
(1)
At temperatures higher then 70 C the FID of homopolymer iPP samples was fitted with a linear combination of one Abragam, one Gaussian and one exponential function, i.e.,
A(t ) = A(0) rigid exp[(t / T2rigid ) 2 ] [sin (at ) / at ] + A(0) semirigid exp[(t / T2semirigidt ) 2 ] + A(0) soft exp[(t / T2soft )]
(2)
The parameter a is related to the second and fourth van Vleck moments. The transverse relaxation times (T2), which are characteristic of different slopes in the magnetization decay curve, are related to the mobility in each fraction. The relative fractions of the relaxation components, {A(0)k/[A(0)rigid + A(0)semi-rigid + A(0)soft]} x 100%, represent the relative amounts of hydrogen atoms (mass fractions) of iPP phases/fractions with different molecular mobility. Repeated experiments for the same sample indicated that the relative error of the extracted relaxation parameters was about 1 %.
67
modulus [N/mm ]
1600
1500
1400
Figure 6.1. Flexural modulus as a function of aging time at 23 C for homopolymer iPP sample.
3.0
parallel perpendicular
2.5
2.0
100
200
300
time [hours]
Figure 6.2. Impact strength in parallel and perpendicular directions at 23 C as a function of aging time for a homopolymeric iPP sample.
1.32 1.30 shrinkage [%] 1.28 1.26 1.24 1.22 0 100 200 300
parallel perpendicular
time [hours]
Figure 6.3. Relative shrinkage in parallel and perpendicular direction versus aging time for homopolymer iPP sample.
68
The decrease in impact strength in both directions (parallel and perpendicular), measured at 23 C with aging time can be observed in Fig. 6.2. Free shrinkage experiments were performed to investigate the changes of the organization of the amorphous phase depending of the aging time. Fig. 6.3 shows the change of the degree of shrinkage in time. In general, the shrinkage behaviour is non-linear and increases with increasing aging time.
600
I(2)
400 200 0 0 5 10 15 20 25 30 35
2
Figure 6.4. X-ray patterns of both initial and aged homopolymeric iPP samples
temperatures. This pre-treatment can be compared with the thermal treatment used to measure the PP samples by solid-state NMR. After this isothermal treatment, the samples were cooled with 10 C/min to 0 C, where they were kept for 5 minutes to equilibrate and then heated at 10 C/min to 200 C (~ 1st heating curve), kept for 5 minutes at this temperature and cooled at 10 C/min to 0 C (~ cooling curve), after equilibration time of 5 minutes at 0 C the samples were finally reheated at 10 C/min to 200 C (2nd heating curve). The so-called 1st heating curve provides information concerning the thermal treatment (a certain time at a certain temperature), whereas both the cooling curve as the 2nd heating curve are being used to verify whether or not the thermal treatment has influenced the thermal response of the material, i.e. molecular and/or morphological changes.
Isothermal Measurements. The isothermal measurements have been analyzed. No crystallization phenomena could be detected. Although crystallization is present, the DSC does not record a crystallization signal during the isothermal treatment. The crystallization is that slow, it can not be seen due to the noise of the baseline (signal/noise ratio too small). Dynamic Measurements. An offset and melting temperature, Tm, can be calculated by the 1st heating curve. All results are presented in Table 6.1. In order to compare the thermal pretreated samples with non-treated PP, offset and apparent offset temperatures are being used based on the previous article [Hed2007]. The crystallization temperature, Tc, the enthalpy of crystallization, Hc and the melting temperature, Tm (2nd heating curve), are also presented in Table 6.1. The enthalpy of the cooling curve is being used to verify the reproducibility of the enthalpy of all samples. Table 6.1 shows that by calculating the mean crystallization temperature, Tc, average = 121.0 0.3 C, the mean crystallization enthalpy, Hc, average = 98 1 J/g and the mean Tm (= 164.0 0.6 C ) one may conclude that the thermal treatment does not change the thermal response of the material after erasing the thermal history. The thermal treatment is a reversible process, which means that no structural changes took place within the homopolymeric iPP sample.
70
Table 6.1. Changes in offset temperature Toffset, crystallization temperature Tc, the enthalpy of crystallization Hc, and melting temperature Tm of homopolymeric iPP samples as a function of aging time and aging temperature Dynamic measurements Thermal treatment Time (min) 5 10 60 120 480 1440 10 60 120 480 1440 10 60 120 480 1440 Average STDEV 28 70 130 Temp (o C) 0 Toffset (o C) / 132 133.4 133.7 134.2 134.9 74.4 75.7 78.3 77.1 80.5 31.4 34 35.8 36.9 40.2 Tm1stH (o C) 162.9 164.1 164.9 164.8 164 164.8 164.2 163.8 163 162.9 164.2 163.9 164.7 164.2 163.3 162.8 163.9 0.7 Tc (o C) 121.2 120.9 121.3 121.1 120.5 120.6 120.9 121.4 120.7 120.7 120.8 121.1 121.1 121.1 121.4 121.5 121 0.3 110-130o C
Hc (J/g)
100 98 97 97 98 98 98 98 99 99 97 98 97 98 98 98 98 1
Tm2stH (o C) 163.3 164.2 164.7 164 163.2 163.7 164.4 164.4 164.2 163.2 164.3 164.2 165 164.5 163.8 164 164.0 0.5
For the samples kept at respectively 28 and 70 C, it is very difficult to determine the Toffset due to the small Cp step and the baseline curvature. The offset temperature of the transition related to the thermal treatment seems to shift to higher temperatures when more time is applied at this isothermal temperature (Table 6.1). The offset temperature values are plotted as a function of the log time and the data are linearly fitted (Figure 6.5). The slope of these lines present the variation of the offset temperature with time, which is the largest for the sample kept isothermal at the lowest temperature (28 C). The offset temperature differences between 1440 minutes and 10 minutes at a well-defined temperature (130, 70 and 28 C) are presented in the Table 6.2. The lower the applied isothermal temperature, the larger the difference is of the offset temperature (slope in Figure 6.5). 71
80 60
y = 27.3 + 3.89x
1000
Table 6.2. Offset temperature difference between 1440 minutes and 10 minutes for annealed PP samples at 28, 70, and 130 C. temperature ( C) [1440 10] ( C) 130 70 28 2.9 6.1 8.8
These results clearly indicate long term changes when keeping homopolymeric PP samples for elaborate times at a well defined temperature. The changes are recorded as an increase of the Cp, which may indicate that the amorphous phase is changing rather than the crystalline phase. Most likely the amorphous and rigid-amorphous phases are seriously influenced by time and temperature effects. The question rises to which extent these long term changes influence the crystalline and/or the amorphous phase.
annealing
Figure 6.6. Changes in the morphology of homopolymeric iPP samples due to aging/annealing.
72
Figure 6.6 presents a hypothesis how annealing influences the amorphous (La), rigidamorphous (Lr-a) and crystalline phases (Lc). Rigid-amorphous phase increases, Lr-a++, amorphous phase decreases, La--, and the crystalline phase increases slightly, Lc+ (3% of crystallinity increase), with time/temperature treatment. These results are in line with the observed NMR results.
6.7 Changes induced by aging at 28 C in the phase composition and chain mobility of homopolymeric iPP samples
Analyses of FID of injection-molded, crystallized form the melt and quenched homopolymeric iPP samples, in terms of rigid and soft fractions were performed at a temperature above Tg (0 C) of 28 C. At this temperature we are claiming that the contrast in chain mobility of the rigid and soft fractions of iPP is enhanced, and the aging process at this temperature can be followed over time with excellent sensitivity. The high amount of the rigid fraction detected by NMR at 28 C is due to the contribution of the immobilized soft fraction to the amount of the crystalline phase. Aging at 28 C for 1000 hours causes 1 2 wt% increase in the amount of the rigid fraction at the expense of the soft fraction (Fig. 6.7a).
a) 90 85 content [%] 80 20 15 10
quenched rigid fraction soft fraction injection molded rigid fraction soft fraction melt-crystallized rigid fraction soft fraction
73
Figure 6.7. (a) The amounts of rigid and soft fraction of injection molded, crystallized from the melt and quenched homopolymeric iPP samples as a function of the aging time at 28 C. (b) The effect of aging at 28 C on chain mobility, as determined by the T2 relaxation time for the different fractions of injection molded, crystallized from the melt, and quenched samples.
For all homopolymeric iPP samples the recorded amounts of the rigid phase increased systematically with increasing aging time (Fig. 6.7a). The systematic changes in the phase composition show that the aging persists for long times with no evidence of reaching an equilibrium state. The chain mobility of the rigid fractions ( T2rigid ) of iPP is affected by aging at 28 C during 1000 hours, whereas the chain mobility in the soft fraction ( T2soft ) decreases over time (Fig.6.7b). The decrease in the chain mobility of the soft fraction due to the aging was also observed by Agarwal [Aga1962] and VanderHart [Van2003]. To calculate a relative rates of aging, the amounts of the rigid and soft fractions, and the chain mobility ( 1 / T2rigid , 1 / T2soft ) data were plotted against aging time on a log scale, and a linear fit was made to the data. The changes in the phase content and chain mobility due to the aging at 28 C are approximately linear on a logarithmic scale over a time of 1000 hours. The overall slopes observed in Fig. 6.7a on a log time scale are not much different for the injection molded (slope = 0.53), melt-crystallized (slope=0.42) and quenched (slope = 1.2) samples. In the case of chain mobility in soft fractions the difference in the slopes are bigger (Fig. 6.7b).
74
6.8 Changes induced by aging at 70 130 C in the phase composition and chain mobility of homopolymer iPP samples
At temperatures between 70 C to 130 C the proton NMR free induction decay (FID) measured for homopolymeric iPP samples crystallized from the melt is decomposed into three components assigned to three fractions with different chain mobilities, manifested in the transverse magnetization times, i.e., the rigid ( T2rigid ), semi-rigid ( T2semi rigid ), and soft ( T2soft ) fractions. At this temperature the contrast in the chain mobility of the rigid, semi-rigid and soft fractions of all homopolymeric iPP samples is enhanced and a quantitative analysis can be performed. Aging in the temperature range from 70 C to 130 C for 150 hours causes 1 2 wt% increase in the amount of the rigid fraction at the expense of the soft fraction (Fig. 6.8a). The amount of the semi-rigid fraction remains almost constant in all the cases (Fig. 6.8b). For all aging temperatures a decrease in the chain mobility of semi-rigid and soft fractions is observed. The chain mobility in the rigid fraction slightly decreases (Fig. 6.9).
a) 80 rigid fraction [%] aging temperature 70 C 110 C 120 C 130 C
70
60
15
14
13
75
20
15
Figure 6.8. The amounts of rigid (a), semi-rigid (b), and soft (c) fractions of homopolymeric iPP samples crystallized from the melt as a function of the aging time for different aging temperatures.
a) 7.2x10 7.1x10
rigid -2
-2
-2 -2
1/T2
-2
-2
2.1x10
semi-rigid
-2
-2
1/T2
-2
-2
76
1/T2
Figure 6.9. The effect of different aging temperatures on the chain mobility in the (a) rigid (T2rigid), (b) semirigid (T2semi-rigid), and (c) soft (T2soft) fractions of melt-crystallized homopolymeric iPP samples.
Using the same approach as in the previous paragraph it is possible to calculate the rates of aging, at different aging temperatures, in terms of the slopes of plots of the amounts of the rigid and soft fractions versus log annealing time and from the plots of T2rigid , T2semi rigid , and T2soft versus log annealing time. In a good approximation all the plots yield a linear relationship at long aging time. The results obtained from those plots are summarized in Table 6.3 and Table 6.4. Table 6.3. Overall slopes estimated from fitting of the dependences of the rigid and soft fractions on the logarithm of aging time at different aging temperatures. aging temperature 70 C 110 C 120 C 130 C slope rigid fraction 0.64 1.56 1.64 2.60 soft fraction -0.55 -1.24 -1.27 -1.7
Table 6.4. Overall slopes estimated from fitting of the dependences of the 1/T2rigid and 1/T2soft on the logarithm of aging time at different aging temperatures. aging temperature 70 C 110 C 120 C 130 C slope rigid mobility -0.04 -0.01 -0.02 -0.01 soft mobility -14.01 -14.22 -14.51 -28.63
77
6.9 Discussion: physical aging in homopolymeric iPP samples by NMR and mechanical analysis
[Hed2007] has reported that the morphology of isotactic polypropylene at temperatures well above the glass transition can be described by a three phases model of rigid, semi-rigid and soft phases. At temperatures close to 25 C the morphology of iPP can be described by a twophase model of rigid and soft fractions. The amounts and chain mobility of each fraction are important parameters regarding the physical and mechanical properties of iPP. Those quantities typically change when the sample is exposed to temperatures well above the glass transition due to aging and annealing [Hik1997, Ala1999, Aki1999, Mar2000, Iji2000, Mai2000, Weg2000, Ali2001, Lab2001]. It is known that the aging of iPP samples after the fabrication causes changes in the structure and properties of the materials. This process has been employed to improve the final properties of the polymers via healing of defects and relaxing residual stress and strain. The observed changes in the mechanical properties (Fig. 6.1, 6.2, 6.3) with the aging time indicate that micro-structural reorganization take place within the material. An increase in elastic modulus has been observed (Fig. 6.1), while the impact strength decreases (Fig. 6.2) with increasing aging time. The decrease in the impact strength is associated with an increasing the amount of the rigid fraction during aging. Property changes occurring upon aging are generally attributed to the relaxation process in the amorphous phase, leading to a reduction in free volume (Fig. 6.3), which will be recovered at higher temperatures, implying that aging is a fully thermoreversible process. The mechanical data imply that aging restricts the motion of the polymer chains, and a good way of probing such changes is to study isothermally the changes in the NMR signal for a long period of time. In order to follow with high sensitivity the changes in the structure at 28 C the amounts and the chain mobility of the rigid and soft fractions are plotted versus log annealing
time. When the sample is exposed at temperature of 28 C for 1000 hours after preparation the
amount of the rigid fraction gradually increase in time at the expense of the soft fraction (Fig. 6.7ab). In contrast to the characteristic values describing the mechanical behaviour, the increase in the amount of the rigid fraction is not very pronounced and is not sufficient to attribute the significant enhancement of mechanical properties during aging just to the secondary crystallization. The chain mobility of the rigid fraction remains almost constant and the chain mobility of the soft fraction decreases with time.
78
Based on information from Fig. 6.7b, it is plausible to conclude that the aging at 28o C influences mainly the soft fraction of all homopolymer iPP samples, and the soft fraction becomes more restricted possibly due to molecular ordering. Our observations are in line with Struiks model [Str1987]. The information provided by Fig. 6.7(a) and (b), with the available level of resolution, indicates that the aging rate is higher in the case of injection moulded and quenched homopolymeric iPP compared to the material crystallized from the melt. From these results we can conclude that important factors determining aging are the morphology of the sample, and a relationship between the physical aging kinetics and the relative amount of the soft fraction. At 28 C we cannot conclude about the changes in the semi-rigid fraction due to aging, because at this temperature the NMR signal can be decomposed into only two components. In order to study the changes in the semi-rigid fraction due to aging, the temperature of the aging and NMR experiment was set within 70 C to 130 C. Increasing the aging temperature from 28 C to 70 - 130 C allows molecular motions to occur, which may lead to structural rearrangements and the observation of the phenomenon of physical aging. According to the results shown in Figure 6.5, Figure 6.8 and Figure 6.9, it is plausible to conclude that in homopolymeric iPP samples crystallized from the melt physical aging influences both the semi-rigid and soft fractions. These changes can be due to the small crystals formed by secondary-crystallization, which hinder the mobility of the soft fraction. The NMR data are consistent with the possibility of the formation of the small crystals, but do not prove this type of structure. Based on the chain mobility considerations, the formation of the new crystals, due to the secondary crystallization, takes place mainly in the semi-rigid and soft fractions, where the chain mobility decreases with increasing the aging time. No evidence of formation of thinner lamellae growing between the primary one due to aging is pointed out by X-ray data (Figure 6.4) Based on the results summarized in Table 6.3 and Table 6.4 the aging rates increase with increasing aging temperature. The combination of results from NMR, DSC, X-ray and mechanical properties as a function of the age of the polymer on the basis of the current understanding of the dependence of those parameters on the polymer structure suggest an explicit interpretation of the changes in the structure, which accounts for the variation of the properties during aging. The structural changes are considered to be associated with a molecular rearrangement generally similar to secondary crystallization accompanied by a decrease in segmental chain mobility of the amorphous chains involved in the diffusion process. A decrease in the segmental chain mobility, by increasing the structural packing density in the amorphous or intercrystalline 79
regions due to the aging, results in an increase in the elastic modulus (see Fig. 6.1.). Tobolsky [Tob1965] states that for polypropylene and another semi-crystalline polymers, the crystalline regions act in part like crosslinks, thereby tending to increase the modulus by increasing the resiliency contributions of the amorphous or semi-rigid regions. The effect of the formations of small crystals due to secondary crystallization may therefore be viewed as effectively increasing the modulus.
6.10 Conclusions
In this chapter, we have investigated the influence of morphology, aging time, and aging temperature on the physical aging behaviour of homopolymeric iPP samples. We characterized the morphology in terms of the two- and three-phase model comprising rigid and soft phases, and rigid, semi-rigid and soft phases, respectively. Injection molded, crystallized from the melt, and quenched homopolymeric iPP samples provided us with specimens of widely differing morphologies. Crystallization from the melt resulted in a relatively lower amount of the soft fraction. This is consistent with a lower degree of constraints imposed by the crystallites on the mobility of the noncrystalline chain segments when crystallization occurs under conditions of high segmental mobility. The extent and kinetics of physical aging in these samples were characterized following the changing in the phase content and chain mobility with time. These changes in the amounts of the rigid, semi-rigid, and soft fractions, and chain mobility in each fraction ( 1 / T2rigid , 1 / T2semi rigid and 1 / T2soft ) displayed a linear dependence on log(aging time). The changes in the phase content and chain mobility with time indicated that the extent of physical aging in semi-crystalline homopolymer iPP samples is significant and easily quantifiable. As a result, physical aging must be considered as an important factor when semi-crystalline iPP samples are designed for use in applications that involve prolonged exposure to aboveambient temperatures. We have observed that the chain relaxation rate in semi-crystalline homopolymeric iPP is sensitive to the relative amounts of the semi-rigid and soft factions. Specifically, for a given aging history, the chain relaxation rate was observed to increase in the samples with a more stable thermodynamic morphology. This observation raises the possibility that the aging takes place primarily in regions of highest mobility that are more distant from the fold surface of the primary crystallites. The observation reported here also suggests that the conformational entropic state of the semi-rigid and soft fractions is more relevant to physical
80
aging in semi-crystalline polymers than the proximity of the aging temperature to their respective glass transition temperatures.
81
82
Chapter 7
Uniaxially Deformed Isotactic Polypropylene
7.1 Introduction and motivation
The mechanical properties of semi-crystalline polymers are strongly affected by changes in the degree of crystallinity due to variations in the molecular structure, different processing conditions and loading conditions [Pop1987, Hal1972, Boy1979, Boy1983, Str1980, Man1964, Man1985, Neg1992, Wor1996, Boy1984, Boy1985]. Loading conditions include the drawing temperature, draw ratio, and the strain rate [Bel1993, Gos1996, Sil1998]. The deformation process of semi-crystalline polymers has been extensively investigated for many years using a variety of methods [Fis1962, Pet1967, Hay1965, Pet1967, Pet1969, Fis1969, Ste1969]. When semi-crystalline polymers are highly drawn, the shear forces destroy the spherulites. This is followed by the reorganization of the spherulitic morphology to another type of morphology depending on the material drawn and the drawing conditions. Several modes of deformation above glass transition temperature have been identified in semicrystalline polymers. According to Flory and Yoon [Flo1978], the plastic deformation occurs by partial melting and re-crystallization of the crystalline phase. However, according to Peterlin [Pet1973] the plastic deformation during the application of stress occurs in three stages. In the first stage, the plastic deformation of the original spherulitic structure occurs. The deformation occurs by strain of the interlamellar soft regions with a combination of interlamellar separation, lamellae stack rotation, interlamellar shear deformation and cavitation [Li1999, Lin1974, Zha2000]. In the second stage, the discontinuous transformation of the spherulitic structure to a fibrillar morphology takes place. During this stage, the lamellae in micronecks are fragmented and the broken crystal blocks are incorporated into newly formed microfibrils. In uniaxial drawing, the blocks are not only oriented into the drawing direction, but also in the direction perpendicular to it [Bra1986]. The chains connecting the broken crystal blocks in the lamellae serve as intramicrofibrillar tie molecules. In the third stage, the plastic deformation of the fibrilar structure occurs. The drawing of the microfibrillar structure [Pet1973] is completely different from the initial plastic deformation that transforms the lamellae into the microfibrillar morphology. The thickness of the fibrils changes almost proportionally to the draw ratio, as a consequence of the much smaller shear
83
displacement of the constituting microfibrils, caused by the shear stress on the fibrils [Pet 1987]. Crystallinity, molecular mobility, lamellar thickness and structural orientation in both soft and crystalline phases are important parameters regarding the physical and mechanical properties of stretched and non-stretched iPP. There are several methods for the structural characterization of the soft and crystalline phases of the polymers. One of the methods - solidstate NMR - is capable of characterizing the crystallinity, the molecular mobility, domain thickness and specifically the orientation distribution function of semi-crystalline polymers [Sch1994, Sch1991, Gun1992, Kiv1986, Bud2003, Ba1998, Che2005, Gol1961, Dem1996]. Another method is IR spectroscopy, widely used to determine the degree of crystallinity of iPP [Hei1959, Sam1981, Kis1983]. Knowledge of strain-induced order in the crystalline and soft phases of stretched iPP under various conditions is valuable information for understanding the structure reorganization mechanisms, and their relation to the mechanical properties of iPP. Wide-angle X-ray diffraction (WAXD) provides information about the crystal lattice order, strain-induced changes in orientation of crystalline domains, and highly oriented crystalline structures in the fibrilar morphology generated in the neck part of uniaxially stretched polymers [Men2003]. Very little information about strain-induced changes in the soft component can be obtained from WAXD. Small angle X-ray scattering (SAXS) provides information about lamellae packing, orientation, and long period [Zhu2004, Wen1993, Zhu2005, Mez1994, Aur2001]. The spectroscopic evidence from infrared (IR) and NMR for structural reorganization during deformation is rather indirect as far as morphology is concerned. Transmission electron microscopy (TEM) yields direct access to the details of the lamellar morphology [Yam1998, Zho2005]. It is thought that IR and NMR are complementary techniques providing morphological evidence on nanometer scale phenomena for stretched iPP, the phase composition, molecular mobility and the orientation. In a recent investigation on iPP by NMR, SAXS, WAXD, DSC and TEM [Hed2007] the phase composition, chain mobility, and domain thicknesses were correlated. The focus of this chapter is to study the change in the phase composition/crystallinity, molecular mobility, and lamellae thicknesses in iPP upon uniaxial drawing at different drawing temperatures and rates. Two complementary methods, 1H solid-state NMR and IR, are used. The study largely focuses on strain-induced changes in the soft phase that have not been analyzed previously in detail. 84
7.2 Materials
The isotactic polypropylene (iPP-PP575P) homopolymer samples that were used for the present study were manufactured by SABIC Europe BV. The material is characterized by a melt flow index of 10g/10 min. It was injected moulded into plates of 12 cm x12 cm and 4 mm thickness using an Engel 45A machine. The melt temperature was set to 235 C, the holding pressure, the holding time, the cooling time and the overall cycle time were 40 MPa, 20 s, 20 s and 49.5 s, respectively. Dog - bone shaped tensile bars with a length of 70 mm, thickness of 4 mm and gauge dimensions of 20 mm length were cut from the injection moulded iPP plates. The samples were uniaxially stretched at different temperatures (25 C, 80 C, and 110 C) and different rates (1, 10, 50, 100 mm/min) to draw ratios in the range of = 7 16 using a Zwick Z050 tensile machine. Based on the literature information [Pet1971] we assume that in this deformation range the spherulitic structure is transformed into fibrillar structure. Since no homogeneous neck formation was observed at = 1 6, samples in this range of strain were not studied. All the NMR and IR experiments in this study were performed on the neck part of stretched samples after 2 hours of strain recovery. New tensile bars were used for each NMR experiment as a function of temperature, strain and deformation rate.
drawing temperatures, the increase in the chain mobility of iPP chains in the soft and crystalline regions (see below 7.5) results in a lower stress required for strain-induced morphological changes. Strain hardening is related to an increase in stress when the neck propagation over the whole length of the gauge part of the sample is achieved. The onset strain hardening seems to be slightly shifted to higher draw ratio with increasing temperature.
b)
100 80
stress [MPa]
60 40 20
stress [MPa]
80 40 0
0 1.0
1.2
1.4
12
16
1.0
1.2
draw ratio
10
Figure 7.1. Stress-strain curve recorded (a) at 25o C, 80o C and 110o C at a constant drawing rate of 10 mm/min, and (b) with different drawing rates of 1, 10, 50, and 100 mm/min at 80 C. The hashed area corresponds to the draw ratio where the NMR experiments were performed for recovered after strain sample.
Figure 7.1b shows the stress-strain curves of iPP samples stretched at 80C with different drawing rates of 1, 10, 50, and 100 mm/min. Both, the yield stress and the strain hardening increase with increasing strain rate at 80 C (see Figure 7.1b). These changes indicate that chain mobility plays an important role in the molecular rearrangements that are required for strain-induced morphological changes. Therefore, the strain increases with the deformation rate.
7.4 Crystallinity by IR
Infrared analyses were carried out on the samples draw to =7, 10 and 12.5 at 80 C and 10 mm/min. Dichroic and non-polarised spectra of one of the samples (=10) are provided in Figure 7.2. It is clear that orientation of the sample has a distinct effect on the acquired spectra, indicating that the sample is non isotropic. The peak heights of the relevant peaks were extracted from these spectra based on the definitions mentioned in the experimental section [Kis1983] and used to calculate the crystallinity (cf. Table 7.1). Due to the 86
experimental errors we cannot draw a conclusion about the dependence of the crystallinity on the draw ratio. However it is clear that the value of rigid fraction from NMR is in the same order of magnitude as the crystallinity measured by IR.
Table 7.1. Crystallinity of uniaxially stretched iPP determined by IR spectroscopy using the Kissin [Kis1983] method. The samples were uniaxially drawn at 80 C with 10 mm/min.
Sample = 7 = 10 = 12.5
a)
Crystallinitya 67 66 69
2.8
Amplitude signal [arb. units]
2.0
1.2
Figure 7.2. IR spectra of the iPP drawn to =10 at 80C acquired with and Polarized and non-polarised lighter. The drawing rate was 10 mm/min.
decreases from 120 C onward due to the partial melting. In the temperature range from 40 C to 60 C, a significant fraction of the soft phase is immobilized and contributes to the amount of the detected rigid fraction (see Figure 7.3a). In this temperature range, the amount of the rigid fraction is composed of the crystalline phase and immobilized fraction of the soft phase. The amount of the soft fraction increases upon increasing the temperature. This increase is accompanied by an increase in molecular mobility of the less constrained chain fragments in the soft phase as it follows from T2soft. In the temperature range from approximately 60 C to 100 C the amount of the rigid fraction is close to the crystallinity determined by SAXS and DSC [Hed2007]. Above 110 C, the amount of the rigid fraction of iPP decreases due to the partial melting of thin lamellae. The iPP samples, having recovered after strain show similar temperature dependence of the amount of each fraction as non-deformed iPP samples.
a)
=12.5
80 60 40 20 0 b)
content [%]
40
T2 semi-rigid T2 soft
=7 T2 rigid
T2 [s]
100
T2 semi-rigid T2 soft
=1 T2 rigid
T2 semi-rigid T2 soft
Figure 7.3. (a) Temperature dependence of the amounts of rigid, semi-rigid, and soft fractions of non-stretched iPP (=1) and uniaxially stretched iPP (=7 and 12.5). The samples was stretched at 80 C with a draw rate of 10 mm/min. (b) Temperature dependence of the 1H T2 relaxation times (T2rigid, T2semi-rigid and T2soft) for both nonstretched and uniaxially stretched iPP samples. The NMR experiments were performed at 70 C.
88
Thus, above 40 C, three different fractions (Figure 7.3a) with different mobilities (Figure 7.3b) are detected. The characteristic decay time T2 is related to the chain mobility in different fractions. The relaxation time (T2rigid) of the rigid fraction increases slightly with temperature in all the samples. The small increases in T2rigid with increasing temperature could be related to c relaxation process [Sch1991]. The intermediate T2semi-rigid relaxation time increases with increasing temperature, but its value essentially does not exceed T2rigid for the same sample at a given temperature. This indicates that the segmental motion responsible for
T2semi-rigid is largely restricted. A significant difference in temperature dependence for the
sample with different draw ratios is observed for the relaxation time T2soft of the soft fraction (Figure 7.3b). With increasing draw ratio the constraints on the segmental mobility in both soft and semi-rigid component increases. This leads to a decrease of T2semi-rigid and T2soft, is especially in the strain-hardening region ( 7) of the samples.
7.6 Effect of drawing ratio on the phase composition and chain mobility
In order to study the structural changes caused by drawing, the amounts of rigid, semi-rigid, and soft fractions that was measured at 70 C is plotted as a function of the draw ratio for iPP samples which were uniaxially stretched at 80 C, with a drawing rate of 10 mm/min (Figure 7.4a). Drawing to = 7 cause a small increase in the amount of the rigid fraction, which indicates the presence of highly strained tie molecules. With the draw ratio increasing to 10 16, a further increase in the amount of the rigid fraction is observed. This increase can be attributed to strain-induced crystallization due to high straining of chains in the soft component and/or large immobilization of chains in the soft phase due to their elongation. In the deformation range from =7 to =12.5 the amount of the rigid fraction is in good correspondence with the degree of crystallinity obtained by IR (see Figure 7.4a). Thus, the increase in the amount of the rigid fraction is mainly due to strain induced crystallization.
89
a) 80 content [%] 60 40 20 0 0
b) 250 T2 200
rigid semi-rigid soft crystallinity/IR
rigid semi-rigid soft
T2 T2
T2 [s]
150 100 50 0 0 4
8 12 draw ratio
16
8 12 draw ratio
16
Figure 7.4. (a) Dependence of the amounts of rigid, semi-rigid, and soft fractions at 70 C on the draw ratio of iPP uniaxially stretched at 80 C (=716). The drawing rate was 10 mm/min. The amount of the rigid fraction () determined by NMR at 70 C is compared with the degree of crystallinity () determined by IR. (b) Draw ratio dependence of the 1H T2 relaxation times for both non-stretched and stretched iPP samples. The T2 values are assigned to the rigid, semi-rigid, and soft fractions of iPP as shown in the figure.
In order to reveal the effect of drawing on the mobility of the rigid, semi-rigid and soft fractions, T2rigid, T2semi-rigid and T2soft are shown as a function of draw ratio in Figure 7.4b. At = 7, both T2semi-rigid and T2soft are significantly shorter than for non-stretched iPP, indicating that the molecular mobility of the semi-rigid and soft fractions largely decreases upon drawing. At >7, T2semi-rigid is essentially unaffected by strain, while T2soft continuously decreases up to the highest draw ratio (=16). The semi-rigid fraction is not affected any more by deformation and the structural changes of the soft phase continue. All these results regarding mobility considerations have important implications for the understanding of the deformation mechanism, and the characteristic changes of the tensile modulus with the draw ratio, in terms of the corresponding changes in the morphology.
7.7 Effect of drawing rate on the phase composition and chain mobility
The effect of the drawing rate on the phase composition and molecular mobility is investigated for iPP samples uniaxially stretched at 80o C to a draw ratio = 10 with drawing rate d / dt of 1, 10, 50 and 100 mm/min. The NMR experiments were performed at 70 C. The amounts of the rigid, semi-rigid, and soft fractions and the chain mobility in each fraction of uniaxially stretched iPP are hardly affected by the drawing rate within the limit of the 90
experimental errors (Figure 7.5 a and b). These results show that the changes in the morphology and chain rearrangements upon uniaxially drawing of iPP are fast compared to the total time scale of explored deformation. Another explanation can be the relaxation process which occurs when removing the sample from the tensile machine.
a)
b)
60
60 content [%] 40 20 0
T2 [s]
50 40 30 20 10
T2 T2 T2
Figure 7.5. (a) Dependence of the amounts of rigid, semi-rigid, soft fractions on the drawing rate of iPP uniaxially stretched at 80 C and = 10. (b) Proton T2 relaxation times for uniaxially stretched iPP samples as a function of drawing rate. The NMR experiments were performed at 70 C. The lines are guides for the eyes.
7.8 Effect
of
drawing
temperature
on
the
phase
functions of both drawing temperature and drawing ratio. The amount of the rigid fraction increases with increasing drawing temperature for all draw ratios at the expense of both the semi-rigid and the soft fractions. The chain mobility in all fractions is lower at higher deformation temperatures. The observed changes suggest that the -relaxation influences morphological transformations via shearing and fragmentation of the lamellae upon deformation. The relative difference in imperfections of the crystalline structures can be identified by comparing the relaxation time T2rigid. T2rigid value is longer at lower deformation temperatures 91
at the same draw ratio . This indicates larger imperfections of the crystalline phase of iPP stretched at 25 C, as compared to the undeformed sample (Figure 7.6b).
a)
80 rigid o 25 C o 80 C o 110 C semi-rigid o 25 C o 80 C o 110 C soft o 25 C o 80 C o 110 C 0 4 8 12 16
60
content [%]
40
20
b)
draw ratio
T2 rigid
8 12 draw ratio
16
Figure 7.6. (a) Dependence of the amounts of rigid, semi-rigid, and soft fractions on drawing temperature and draw ratios for iPP uniaxially drawn at 25 C, 80 C, and 110 C. A drawing rate of 10 mm/min was used. The NMR measurements were performed at 70 C. (b) Dependence of the 1H T2 relaxation times on the drawing temperature and the draw ratio for iPP uniaxially stretched at different temperatures. The lines are guides for the eyes.
A small increase in the amount of the rigid fraction is observed at strain hardening deformations (>7) that can be due to a large extension of the tie chains. These results suggest that at higher deformation temperatures, higher strain-induced chain orientation and better ordered crystalline structures are formed upon deformation. This is because higher chain mobility facilitates the morphological transformations, and possibly partial melting and straininduced crystallization occur at higher temperatures. The NMR measurements reveal that the drawing temperature has an important influence on the phase composition and molecular mobility of drawn iPP. 92
7.9 The effect of draw ratio and drawing temperature on the domain sizes of iPP
The long period Lp and the thickness of the rigid (dr) and semi rigid + soft (dsrs) domains were studied as a function of both draw ratio and drawing temperature. The values of Lp, dr, and dsrs were determined from the analysis of the spin-diffusion data (measured at 70 C) (Figure 7.7) with the 1D (non-stretched iPP) and 2D (stretched iPP) spin-diffusion models [Hed2007].
1.0
=7
A(t)/A(0)
0.8
0.2
0.0 0
4 td
1/2
8 1/2 [ms ]
12
rigid semi-rigid + soft .......................... =10 rigid semi-rigid + soft .......................... =12.5 rigid semi-rigid + soft
Figure 7.7. Proton spin-diffusion decay and build-up curves for the rigid and semi-rigid fractions of iPP uniaxially stretched to = 7, 10 and 12.5. The drawing temperature and drawing rate were 80 C and 10 mm/min, respectively. The NMR experiments were performed at 70 C. The DQ filter selects the magnetization of the rigid fractions of stretched iPP, as described in ref [Hed2007]. All experiments were performed at low field. The FID was fitted with the sum of one Abragam function and one exponential function. The solid lines represent the best fit to the 2D solutions of the spin-diffusion data for the two-phase morphology.
The different models are required because above yield strain, the spherulitic morphology is transformed into a fibrillar morphology typical for polymeric fibres. The chosen dimensionality of the spin diffusion process is established based on the morphology for non-stretched and stretched polyolefins [Sch1991, Hu2000, Fis1962]. This is also confirmed by the analysis that was outlined in the Chapter 5 describing the morphology of non-stretched iPP [Hed2007], which shows that the dimensionality of the spin-diffusion process is 1D and 2D for initial sample and the uniaxially drawn iPP, respectively. The thickness of the rigid and combined semi-rigid plus soft domains are determined for uniaxially stretched iPP because the small amount of the soft fraction 2-5 wt.-% (Figure 7.4a) does not allow accurate determination of the thickness of this fraction. 93
The long period and the domain thicknesses obtained by the analysis of the spin-diffusion data which are shown in Figure 7.7 are given in Figure 7.8. This figure shows the changes of
Lp, dr, and dsrr upon uniaxial drawing to different draw ratios and at different drawing
temperatures. At a drawing temperature of 25 C, that is below the -relaxation temperature, the long period and the thickness of the rigid domains slightly decrease at the expense of the thickness of the semi-rigid plus soft domains (Figure 7.8). These dependences of the rigid domain thickness dr and long period Lp on the deformation temperature can be explained by brakeage and disordering of crystals. During the uniaxial deformation of iPP at 25 C the probability of the broken crystals to recrystallize and form thicker crystals is small. Straininduced immobilization of the soft phase hinders the displacements of chains in the soft and crystalline phases that are required for lamellae thickening. No significant strain-induced crystallization and/or annealing occur during deformation at 25 C. At higher deformation temperature faster chain mobility (Figure 7.3) can provide the required energy to facilitate the reorganization of chains to cause an increase in the amount of the rigid fraction (Figure 7.6). For higher drawing temperatures (80 C and 110 C), an increase in the long period and thickness of lamellae is observed upon increasing the draw ratio and drawing temperature (Figure 7.8).
dr
16
Lp, dr, dsrs [nm]
25 C
dsrs Lp
12 8 3 2 0
8 12 draw ratio
16
Figure 7.8. Dependence of long period Lp,, the thickness dr of the rigid and semi-rigid + soft (dsrs) domains on the drawing temperature and the draw ratio for iPP stretched to = 7, 10-16 at 25 C, 80 C, and 110 C with 10 mm/min.
The thickness dsrs of the semi-rigid plus soft layer slightly decreases with increasing the draw ratio and drawing temperature suggesting that the crystal thickening is caused by the chain reorganization in the soft phase adjacent to the crystal surface and the annealing process. Thus, deformation at temperatures above 60 C causes an increase in crystal thickness and long period, while deformation at lower temperatures causes a slight decrease in 94
crystal thickness and possibly long period. The main difference between deformation at high and low temperature is attributed to the rate of molecular motion with respect to the rate of molecular displacement with increasing strain.
7.10 Correlation between the amount of the rigid fraction and the thickness of the rigid domains
Figure 7.9 shows the correlations between the amount of the rigid fraction and the domain thickness dr of the rigid phase for all iPP samples uniaxially stretched at different temperatures. For the limited drawing temperature range 80 C - 110 C, the dependence of the amount of the rigid fraction on dr is apparent described by a linear function.
a)
b)
8.9
=7
dr [nm]
drawing temperature 25 C
=10
12.2
dr [nm]
drawing temperature 80 C
=12.5 =10
=14
8.8
8.7 56 58
=12.5
=7
60
62
11.4 56
60
76
dr [nm]
=12.5
13.6
=10
13.2
=7
60
64
80
Figure 7.9. Correlation of the thickness of the rigid domains with the amount of the rigid fraction/crystallinity for iPP uniaxially drawn at 25 C (a), 80 C (b) and 110 C (c) with 10 mm/min.
95
The behavior of the long period presented in Figure 7.10 (see below) is in qualitative agreement with that reported in Ref. [Bal1969]. As proven by SAXS [Bal 1969] for polypropylene, drawing at a temperature higher than 600 C yields an increase of long period whereas deformation below these temperatures causes a decrease of the long period. This is evident from Figure 7.9b and c and Figure 7.9a, respectively. Nevertheless, in our case the correlation is established between the thickness dr of the rigid fraction and the content of the rigid fraction measured by 1H spin diffusion at different drawing temperatures and draw ratios. We shall consider in the following a semiquantitative model to establish the correlation between the thickness dr of the rigid fraction and the content Cr of the rigid fraction in the high drawing temperature regime (Figure 7.9b and c). The changes produced by deformation in polymer morphology are enhanced at higher drawing temperatures. This is due to large mobility of polymer chains at the interface and in the soft fraction. Hence, the thickness of the rigid fraction and its content will increase at higher drawing temperatures. At such conditions the original morphological structure is completely destroyed and every single lamella is broken up into small folded chain blocks. The blocks are then incorporated into long microfibriles. We assume that the spin diffusion process in these microfibriles can be described by a 2D morphology as proposed for polyethylene in Ref. [Hu 2000]. In the approximation of a square section of microfibriles of size dr and length hr the volume is
microfibri Vrigid l d r2 hr . The total volume of the rigid fraction at the end of the drawing process is microfibri Vrigid = N microfiblilsVrigid l N microfiblils d r2 hr . Therefore the rigid fraction content is C r Vrigid ,
correlation between d r and C r (Figure 7.9b and c). Apparently a linear dependence is shown by the data that we believe is due to the limited range of Cr explored in the experiments. Finally, we shall point out that NMR measurements that establish the thickness of the rigid domain and the phase composition are independent.
7.11 Correlation between the long period and the drawing temperatures
The long periods (Lp) were measured for different iPP samples drawn at different draw ratio
(Fig. 7.10). The long period first increases rather slowly with temperature and finally increases rapidly near the melting point. Such investigation was reported before on polyethylene [Bal1969, Pet1965, Cor1967 55].
17 Long period [nm] 16 15 14 13 12 20 40 60 80 100
o
=7 =10 =12.5
120
drawing temperature [ C]
Figure 7.10. The long period of iPP samples uniaxially drawn with 10 mm/min and measured by 1H spindiffusion NMR as a function of the drawing temperatures for different values of draw ratios. The solid lines represent the fit of the data with Eq. (7.1). The fit parameters A and B are reported in Table 7.2.
The long period of drawn iPP as a function of the drawing temperature is shown in Figure 7.10. This dependence can be represented approximately [Bal1969, Pet1965, Cor1967] by the equation:
Lp = A + B , T Td
0 m
(7.1)
0 where Tm is very close to the melting temperature (Tm=160 C) and Td is the temperature of
drawing. The continuous lines in Figure 7.10 represents the fit of the data with Equation. 7.1. The constants A and B are given in Table 7.2 for different draw ratios.
Table 7.2. The values of the constant A and B from Equation (7.2) obtained from a fit of the data shown in Figure 7.10.
They are in the same range of values as those reported for polyethylene [Pet 1965, Cor1967]. The long periods were measured in this case by SAXS. These results prove indirectly that the spin-diffusion measurements give a valid trend for the long periods.
97
One can conclude that the blocks of chains broken in the initial iPP morphology are subjected to such a high concentration of deformational energy that they melt and reorganize according to the temperature of drawing like in polyethylene [Cor1967]. The resultant long period is more characteristic of the drawing temperature than the initial state of the iPP sample. The change of the long period is abrupt, indicating a discontinuous step in the transformation form the original microspherulitic into the fibrillar structure [Bal1969].
7.12 Correlation between the 1H transverse relaxation rate and the domain thickness of the rigid domains
Morphology of semi-crystalline polymers and domain thickness characterize the material on a scale ranging from nanometers to hundreds of nanometers and micrometers. The analysis of the chain dynamics, revealed inter alia by the 1H transverse magnetization rate, is a useful tool to improve the understanding of the microscopic properties of heterogeneous polymers. The value of the 1H effective transverse magnetization relaxation rate (1/T2rigid) for the rigid domains is related to the type and frequency of chain motions, and to intrachain and interchain interactions in those domains. Faster transverse relaxation rates correspond to an increase in the dipolar coupling due to the denser chain packing and/or better crystal organization that could affect chain mobility in the rigid domains. The correlation between 1/T2rigid and domain thickness dr for different draw ratios and drawing temperatures is shown in Figure 7.11. For iPP samples uniaxially stretched at 25 C, the increase in 1/T2rigid and decrease in dr with increasing draw ratio are due to the immobilization of the polymer chains and the fragmentation of the crystalline blocks. At this drawing temperature, the chain mobility is lower and the reorganization of the morphology from spherulitic to fibrillar is restricted. The increase in the values of 1/T2rigid and dr, for iPP samples uniaxially stretched at 80 C and 110 C, are due to a decrease in chain mobility. Moreover, at this deformation temperature the chain mobility is larger than at deformation temperature of 25 C, which facilitates the reorganization of the morphology. The perfection of the crystals can occur during the strain-induced recrystallization process and/or immobilization of the soft phase, which both occur at higher deformation temperatures. We note that all NMR measurements were done at 70 C on stretched iPP samples recovered after strain. Therefore, the dependences shown in Figure 7.11 reflect changes in the thickness and chain reorganization of crystals due to irreversible iPP deformation at different temperatures. 98
a) 8.9
=7
b)
dr [nm]
drawing temperature 25 C
=10
drawing temperature 80 C
dr [nm]
8.8
=7
8.7 0.094
=12.5
0.096
11.4 0.102
0.105
14.0 dr [nm]
=12.5
13.6
=10
13.2
=7
0.105
Figure 7.11. Correlation of the thickness of the rigid domains with the 1H transverse magnetization relaxation rate of the rigid domains (1/T2rigid) for iPP that was uniaxially drawn at 25 C (a), 80 C (b) and 110 C (c) with 10 mm/min.
Elastic recovery in the soft phase can cause some changes in the phase composition, domain sizes and molecular mobility after removing the applied stress. However, these changes would not affect the conclusions about the effects of the draw ratio, drawing rate, and temperature on the structural reorganization during drawing of iPP.
7.13 Conclusion
In this present chapter, we have investigated by 1H solid-state NMR the changes in the phase composition (namely the amounts of rigid fraction/crystallinity, semi-rigid, and soft amorphous fractions), chain mobility, and domain thickness for uniaxially drawn iPP at different draw ratios, drawing rates, and drawing temperatures. The values of the rigid 99
fraction/degree of crystallinity obtained by NMR and IR coincide within a few percent. The small difference can be explained by the analysis of NMR and IR data using three- and twophase model, respectively. An increase in the drawing rate from 1 mm/min to 100 mm/min does not affect the phase composition and chain mobility for iPP stretched at 80o C. The drawing temperature was shown to have a significant effect on the phase composition, chain mobility, long period, and lamellar thickness. The results show that the amount of the rigid fraction/crystallinity increases as the draw ratio and drawing temperature increase for the entire range of draw ratios ( = 1, 7 - 16). For example, the amount of the rigid fraction in non-drawn iPP is 55 wt.-% at 70o C. Its value at the same temperature increases up to 65 wt.% for iPP stretched to =12.5. Higher draw ratio and drawing temperatures above 80 C resulted in an increase of the thickness of the lamellae from 11 nm to 16.5 nm that can be explained by partial melting and recrystallization as well as immobilization of the soft phase. We can conclude that the phase composition, chain mobility and the thickness of domains largely depend on the deformation conditions, namely the draw ratio and the drawing temperature. Correlations between the following molecular parameters were established: (1) the amount of the rigid domains and their thickness; (2) the chain mobility in the rigid domains and the thickness of these domains. A semiquantitative model was developed to describe the correlation between the amount of the rigid fraction and the thickness of the crystalline domain in the range of high drawing temperatures. Moreover, the correlation between the long period and the drawing temperature was analysed based on a phenomenological law applied before on polyethylene [Cor1967]. These relationships differ for iPP uniaxially drawn at 25 C and at higher deformation temperatures 80 C. The results reveal the important role of chain mobility during strain-induced transformation of the spherulitic morphology to a fibrillar one. At 25 C, the rate of molecular motion is not high enough to follow molecular displacements that are caused by macroscopic strain. This is conformed by the fact that the stress-strain dependence at 25 C largely depends on the deformation rate. High local forces that cause crystal slip damage crystalline structures as can be concluded from a small decrease in the lamellae thickness. At the same time the amount of the rigid fraction increases due to the immobilization of highly strained chain fragments in the soft phase. Deformation at temperatures of 80 C and higher, is not accompanied by high local forces due to more intensive chain motion. Moreover, a heat release during drawing causes partial melting followed by recrystallization. This results in thicker crystals and larger long periods. As far as the chain mobility is concerned, it decreases with increasing draw ratio both in the rigid and 100
soft phases. The decrease is more pronounced at higher deformation temperatures, which points out a more perfect structure reorganization due to partial melting followed by recrystallization at higher deformation temperatures.
101
102
Chapter 8
Impact Modified Copolymer Polypropylene
8.1 Introduction and motivation
Polypropylene (PP) is one of the most important commercial polymers available because of its extremely broad spectrum of applications. Homopolymeric PP however is too brittle for applications such as automotive components that require a polymer with high or medium impact resistance. The most widely used method to improve the impact properties of PP is through copolymerization with another -olefin in a multi-stage reactor system or blending with ethylene propylene rubber (EPR), resulting in a system with dispersed ethylene propylene rubber particles (EPR) in a PP matrix [Ros1992, Kar1999]. The impact properties depend on the EPR rubber dispersion, composition and particle sizes and on the matrix crystallinity [Jan1993, Jan1994, Pre2000, Jan1996]. However, such modifications to improve the impact properties are often detrimental to other desired properties. The development of new systems that show an improved balance between properties represent therefore one of the major challenges to the PP industry. For instance, an improved balance between impact strength and stiffness, impact strength and transparency and impact strength and stress whitening is desired and are areas of continuous research. To partially compensate the decrease in stiffness due to the addition of EPR, addition of nucleating agents such as talc and ADK STAB NA-11 is recommended. Addition of nucleating agents increases the temperature of crystallization resulting in a decrease in cycle for injection moulding [Fil1994]. Nucleating agents increase the number of crystallization sites in polymers and reduce the spherulite size. They cause simultaneous growth and thus evenly sized crystals. Smaller evenly sized crystals improve the physical properties of the polymer like transparency and flexural modulus. A detailed knowledge of the phase morphology is a key parameter for improving the mechanical and optical properties, and for obtaining a proper balance between these properties for different applications. A wide range of characterization methods has been applied to study the PP copolymers. A number of these techniques are suited to give information on the morphology of PP copolymers where the EPR domains typically extend form a few hundred nanometers to a few micrometers in size. Scanning electron microscopy (SEM) has been applied to study the EPR distribution in PP copolymers [Gre2002, Goi1992, Goi1993, Poe2000, Kim1998, 103
Cha2002, Cha2002, Cop1987, Pet1996, Bro1997, Kar1979, Ora2001, Bed2000, Yam1999, Ora1982]. TEM is another technique that has been widely applied to study the morphology in PP copolymers [Gre2002, Kim1998, Nor1995, Yok1997, Nys1999, Sta1998, San1986, Bed2002, Tom2000, Kim1996]. The prepared TEM samples show clear contrast between matrix and rubber inclusions. This technique has the advantage that the lamellae structure of the PP matrix as well as the rubber part can be studied [Poe2000, Ora2001, Sta1998, San1986]. Atomic force microscopy (AFM) is a convenient method compared to SEM and TEM in the sense that less sample preparation could be necessary and several investigations have employed AFM for morphology studies of rubber-modified PP [Cha2002, Cha2002, Nys1999, Bed2002, Nys1995, Swa2000,]. AFM directly reveals the presence of the rubber inclusions on the surface. Nuclear magnetic resonance (NMR) relaxation provides an excellent microscopic probe of molecular or segmental motion in PP copolymer systems [Dou1979]. Sensitive indicators of molecular mobility in different fractions are the nuclear spin-lattice relaxation time T1 and spin-spin relaxation time T2. Previously, solid-state NMR relaxation has been used to study the changes in the phase content and the chain mobility in different fractions of polyethylene and polypropylene [Hed2007]. The first purpose of this chapter is to study the morphological features of impact modified PP copolymer by transmission electron microscopy (TEM), scanning electron microscopy (SEM) and atomic force microscopy (AFM). The second purpose is to study the changes in the phase content and chain mobility due to the change in the rubber content and the addition of nucleating agents (talc and ADK STAB NA-11). The third propose is to establish a correlation between the NMR parameters, DSC results, and mechanical data.
8.2 Materials
The impact modified copolymer polypropylene samples used in our experiments were manufactured by SABIC Europe BV. The material was first compounded at temperature between 244 - 262 C in a ZSK3033 extruder than injected in an Angel 45A machine. The melt temperature was set to 235 C, the holding pressure was 40 MPa, the holding time 20 s, the cooling time 20 s, and the overall cycle time 49.5 s. The mold chosen provides an injection molding plate (65 x 65 x3.2 mm) and a tensile bar (dumbbell shape). Three series of five samples were prepared with different rubber contents (from 0% to 33%) and different nucleating agents (Table 8.1). 104
Table 8.1. Listed details of impact modified PP copolymer samples that were examined in this study
Series 1 1 2 3 4 5 % wt RC 0 5 15 25 33 % Talcum % ADK STAB NA-11 Series 2 1 2 3 4 5 % wt RC 0 5 15 25 33 % talcum 0.05 0.05 0.05 0.05 0.05 % ADK STAB NA-11 Series 3 1 2 3 4 5 % wt RC 0 5 15 25 33 % talcum % ADK STAB NA-11 0.05 0.05 0.05 0.05 0.05
The first series of samples (Series 1) is a regular PP impact modified copolymer, the second series (Series 2) is a PP impact modified copolymer with talc nucleating agent and the third series of samples (Series 3) is a PP copolymer modified copolymer with an ADK STAB NA11 nucleating agent, respectively. The injection molding plates were used for the impact test, stiffness text, CLTE, and NMR experiments. The tests on the injection molding plates were performed in two directions: parallel and perpendicular to the injection molding side because the values measured are significantly different for the two directions. The tensile bars were used for the tensile test.
2 m
0.5 m
Figure 8.1. TEM picture of a regular PP impact copolymer sample with 5% of ethylene-propylene rubber (EPR)
105
Figure 8.1 and 8.2 demonstrate that EPR domains can clearly be seen within the PP matrix for both samples containing 5% EPR and 33% EPR, respectively. It was also observed that increasing the rubber content changes only the density and sizes of the rubber particles in the PP matrix. Because the EPR domains are soft relative to the surrounding PP matrix, lighter regions correspond to the PP matrix and darker regions correspond to EPR domains. It can be seen also that some EPR domains are dispersed and nonuniform in the cross section however, most of them can be approximated as having a rather spherical shape. Comparing the images of Fig. 8.1 and Fig. 8.2, it becomes clear that the rubber domains are more numerous for the sample with 33% rubber content. The average size of the rubber domains is between 0.2 m and 2.4 m for the samples with 5% rubber and between 0.2 m and 3.4 m for the samples with 33% rubber. a b
2 m
0.5 m
Figure 8.2. TEM picture of a regular PP impact copolymer sample with 33% of ethylene-propylene rubber (EPR)
Figure 8.1b and Fig. 8.2b further detail the morphology of the rubber domains of the regular impact copolymer PP samples with 5% and 33 % rubber content, respectively. For both samples the rubber domains can be better visualized, the presence of the dark amorphous phase (rubber) and relatively light crystalline phase (ethylene crystals) being clearer. An additional technique used to visualise the overall shape of the dispersed rubber domains in the PP matrix is the AFM Phase Imaging (Figure 8.3). The lighter colour represents the homopolymer PP matrix whereas the darker colour represents the rubber phase. In both Figures 8.2 and 8.3 the presence of ethylene crystals in the rubber phase can be observed, i.e. the lighter regions within the darker ones.
106
Figure 8.3. AFM phase image revealing distribution of rubber particles in an injection moulded plate of PP impact copolymer (33% rubber).
The size of the rubber domains provided by the AFM images is in good agreement with the size provided by TEM, i.e. between 0.2 m and 3.4 m for the 33% rubber content sample. The AFM scans also provided information with respect to the talcum particles, i.e. their flowinduced orientation and the fact that they are preferentially surrounded by the matrix. Based on AFM scans of specific areas, the changes in the shape of the particles can be monitored throughout the samples surface, even in systems containing high amounts of talcum particles (Figure 8.4). These systems were difficult to investigate by TEM.
Talcum particles
Figure 8.4. AFM height (left) and phase (right) images revealing distribution and orientation of rubber and talcum particles in an injection moulded plate of PP copolymer.
107
temperature of crystallization [ C]
140
130
120
110
10
20
30
120
100
80
10
20
30
Figure 8.6. Heat of fusion of both, the homopolymer and all series of the PP modified copolymer versus rubber content
108
Figure 8.6 represents the heat of fusion of the second heating curve versus the rubber content. A decrease in the heat of fusion for all series impact copolymer PP samples can be observed with increasing rubber content. The heat of fusion of PP matrix is constant and independent of the rubber content. An increase in the heat of fusion is observed due to the presence of nucleating agent.
109
40
35 30 25 20 0 10 20 30
2000
1500
1000
500
Figure 8.8. Flexural test of PP copolymer sample with different nucleating agents
A flexural test was conducted to determine the elastic modulus of each of the sample in the series. The principle of the flexural test is to bend the tested specimen in the middle until the fracture of the specimen or until the deformation reaches a predetermined value. During this procedure, the bending force applied to the specimen is measured. The tests have been performed in two directions. Figure 8.8 represents the flexural modulus determined when the bending force was applied parallel to the injection direction in the injection molding process. It was observed that the flexural modulus decreases with increasing rubber content. Addition of the rubber increases the toughness, however, it is associated with a loss in stiffness. The addition of nucleating agents increases the elastic modulus. The decrease in the modulus associated with the increase in the rubber content can be partly prevented by the 110
simple addition of nucleating agents. For example, the modulus of a sample containing 25% of rubber was brought to the value of the modulus corresponding to a sample with 15% rubber and no nucleating agents by adding nucleating agents.
content [%]
ri g i
d
ry bbe + ru e rfac inte
er Int
e fac
er bb Ru +
50 30
T2 [s]
P) (soft P
20 semi-rigid 10 0
soft
Figure 8.9. (a) The amounts of rigid phase, semi-rigid interface, soft fractions and rubber phase of all series of impact modified PP copolymers. (b) The effect of rubber content and nucleating agent on chain mobility, as determined by the T2 relaxation time for the different fractions the copolymer. Larger amplitudes and/or frequencies of molecular motions lead to longer T2 values. The relaxation characteristics were determined by the analysis of the FID, (rigid and semi-rigid fractions), and the Hahn-echo (soft and rubber fraction). The experiments were performed a with low-field NMR spectrometer.
At 70 C the NMR signal (FID) of all series of PP samples showed four different relaxation times (Fig. 8.9b). The shortest relaxation time T2rigid is associated with the chain mobility in the rigid fraction; the intermediate relaxation time T2semi-rigid is associated with the chain mobility in semi-rigid fraction; the long relaxation time T2soft is associated with the 111
chain mobility in soft fraction, and the longest relaxation T2rubber time is associated with the interface and rubber fraction. The chain mobility of the rigid (T2rigid) and semi-rigid (T2semi-rigid) fractions was found unaffected by increasing the rubber content. This shows that the ethylene-propylene rubber is well separated from the rigid and semi-rigid fractions, and cannot be present in those fractions. Similarly, the increase in the chain mobility of the soft fraction (T2soft) of the copolymer was explained by the presence of ethylene-propylene rubber mainly in the more mobile core part of the soft phase. The longest relaxation time (T2rubber) related to the mobility of the rubber fraction was found to increase with increasing the ethylene-propylene rubber content. This is believed to be related to the average size and size distribution of ethylene-propylene rubber domains and the presence of small ethylene crystals in the rubber domains. In small rubber domains the chain mobility of the rubber molecules is more restricted due to the attachment of a large percentage of these molecules to the fraction of the PP chains in the amorphous phase. In larger rubber domains the percentage of the attached rubber molecules is lower and therefore the chain mobility is less restricted. Another explanation for increase in T2rubber with increasing the rubber content can be the better phase separation between the rubber domains and the amorphous PP phase for the sample with the highest amount of rubber. The large difference in the molecular mobility between PP matrix and rubber matrix is due to Tg much lower for ethylene-propylene rubber than for the PP homopolymer. No significant changes in the phase content and chain mobility were detected by NMR for any of the series of impact modified PP copolymer samples upon addition of nucleating agents. The amount of rubber estimated by low-field NMR was found to be in good agreement with the results obtained from liquid-state NMR and the calculated chemical composition.
112
modified PP copolymer, the amounts and the chain mobility of rigid, semi-rigid, soft, and rubber fractions are plotted versus log aging time (Figure 8.10).
a) 70
rigid semi-rigid soft rubber
content [%]
60
10 1
b) 50 40 content [%] 30 20 10 1 10 aging time [hours] 100
100
d)
100
960
rigid semi-rigid soft rubber
958
T2 [s]
100
Figure 8.10. (a,b) The amounts of rigid, semi-rigid, soft and rubber fractions as a function of aging time for regular impact modified PP copolymer with 5wt% and 33 wt% rubber. (c,d) Chain mobility in rigid, semi rigid, soft and rubber fraction of impact modified PPcopolymer with 5wt% and 33 wt% rubber .
Comparing the results shows in Figure 6.7 and 6.8, and the results of Figure 8.10 we concluded that the changes in the phase content and chain mobility of the PP matrix of the impact modified PP copolymers are the same as in the homopolymer PP samples. We also observed that in time the amount of the rubber fraction decreases whereas the chain mobility of the rubber phase increases, which is probably due to the ongoing separation between the rubber domains and the amorphous PP phase. The highest increase in chain mobility of the rubbery phase is observed for the sample with 33 wt% rubber. Based on chain mobility considerations, a better separation between the rubber domains and amorphous phase of PP will cause an increase in the chain mobility of the rubber domains. This can partly explain the results presented in Figure 8.9. 113
70 C
rigid semi-rigid soft interface + rubber
0 C
rigid
interface
rubber
In contrast to the homopolymer PP matrix, the existence of three fractions with different mobilities is revealed at 0 C for the impact modified PP copolymer. Relaxation components were assigned to these three fractions of different molecular mobility, i.e for the PP matrix (T2matrix), interphase (T2interphase), and rubber (T2rubber) phases, respectively (Figure 8.12). It was observed by NMR that the amount of the interphase increases with increasing rubber content (Figure 8.12a). This is in line with the TEM picture which shows that the sample with 33% rubber content contains bigger rubber domains (Figure 8.1 and 8.2). The chain mobility in the PP matrix remains constant however, indicating that the EPR is well separated from the most crystalline part of the PP matrix (Figure 8.12b). It was further observed that the chain mobility of the interface increases with increasing rubber content. This effect can be explained by a higher amount of mobile EPR chains that is present at the interface. On the other hand, the chain mobility of the rubber phase remains almost constant indicating that the EPR chains in the rubber domains have the same mobility. Using these findings at 0 C, we can now explain better why the mobility at 114
70 C increases with increasing amount of rubber (Figure 8.9). We also assume that the ethylene crystals inside the rubber domains play an important role in the mobility of the EPR chains within these domains.
100
120
80 content [%]
T2 [s]
T2 PP matrix
20 10
T2 interphase T2 rubber 40 20
100
Figure 8.12. (a) The amounts of rigid phase, interface and rubber fractions at 0 C of regular impact modified PP copolymer with different rubber amounts (Series 1). (b) Chain mobility of the rigid (T2matrix), interphase (T2interphase), and rubber (T2rubber) fractions at 0 C of the regular impact modified PP copolymer with different rubber contents.
To further verify the accuracy of the data obtained at 0 C, the sum as determined by NMR of the amounts of the interphase and rubber phase was compared with the same data derived from the chemical composition (Figure 8.13). A good correlation between these data was obtained. The above described analysis by low temperature NMR helped to understanding the mechanical behaviour of the impact modified PP copolymer.
Figure 8.13. Comparison between the amounts of the rubber detected by NMR and the amounts of the rubber estimated from the chemical composition.
115
One of the disadvantages of known PP copolymers is their low impact strength at low temperatures. A method to improve the impact strength is to blend the PP with rubber, also known as rubber toughening. Figure 8.14 represent the impact strength versus temperature for regular impact modified PP copolymer with 5 wt% rubber. The letters indicate the different classifications given to the sample. A brittle sample (B) means that the sample break rapidly, whereas a tough sample (T) yields above a certain value of the applied force. In the transition regime, there are brittle-tough samples (B-T) and tough-brittle (T-B) samples. The sample is called brittle-tough when for the five sample tested there are more brittle than tough samples and the other way around for the tough-brittle samples.
30 20
T-B B
10
B B
B-T
20
40
60
80
100
temperature [C]
Figure 8.14. Impact strength of regular PP impact copolymer with 5 wt% rubber at different temperature
The impact strength at low temperatures depends on the behaviour of the rubber at those low temperatures. To study in detail the rubber behaviour with temperature in impact modified PP copolymer samples with 5 wt% rubber content, the NMR experiments were performed from -40 C and up to 70 C. In the temperature range from -40 C to -20 C (regime I in Figure 8.15ab) the NMR signal can be decomposed in two components with different chain mobility, i.e. a rigid component with low chain mobility and a soft component with higher chain mobility.
116
b)
40 soft
Soft, Rubber [%]
soft + semi-rigid
soft + rubber
T2 soft
T2 soft
II
III
II
T2
III
rubber
20
40
60
80
temperature [C]
-40 -20
0 20 40 60 temperature [C]
80
c)
II
III
rigid soft
Figure 8.15. (a) The amounts of soft and rubber phase at different temperatures for regular impact modified PP copolymer with 33% rubber. (b) Chain mobility of the soft (T2soft) and rubber (T2rubber) fractions at different temperatures. (c) Schematic representation of the morphology at different temperatures by NMR.
In the temperature rage from -20 C to 40 C (regime II in Figure 8.15ab) three different components with different relaxation times can be identified, i.e. a rigid component with low chain mobility, a semi-rigid component with intermediate mobility and a soft component with higher chain mobility. Above 60 C (regime III in Figure 8.15ab) the NMR signal can be decomposed in four relaxation components i.e. a rigid component with low chain mobility, a semi-rigid component with intermediate mobility, a soft component with higher chain mobility, and a rubber component with the highest chain mobility. It can be concluded that by increasing the temperature, the presence of the rubber phase in the impact modified PP copolymer becomes more significant because it provides the entire copolymer with a more rubber-like behaviour. As a consequence, the impact strength of the copolymer increases accordingly.
117
8.9 Conclusion
Improving the impact toughness and stiffness of PP has important practical meaning for extending the range of applications. Obviously the impact toughness of PP can be enhanced by rubber inclusion in the PP matrix. DSC, TEM, AFM, and NMR analysis confirmed that the impact modified PP copolymer consists of multiple phases: PP, EPR (crystalline and amorphous), and PE. Because of the addition of rubber in homopolymeric PP, variations of the molecular structure can be induced. The effects of the addition of rubber and the addition of nucleating agents on the different properties are presented in the following Table 8.2. Table 8.2. Effect of addition of rubber and nucleating agent on crystallization temperature, heat of fusion, yield stress, flexural modulus, phase composition and chain mobilities of impact modified PP copolymers. Properties, phase content and Increasing rubber content Adding of nucleating agent chain mobilities Crystallization temperature No effect Increase Heat of fusion Yield stress Flexural modulus Toughness Rigid fraction Semi-rigid fraction Soft fraction Rubber fraction Mobility in rigid fraction Mobility in semi-rigid fraction Mobility in soft fraction Mobility in rubber fraction Decrease Decrease Decrease Increase Decrease Decrease Decrease Increase No effect No effect Increase Increase Increase Increase Increase Decrease Increase No effect No effect No effect No effect No effect No effect No effect
118
Chapter 9
General Conclusions
Solid state Nuclear Magnetic Resonance has successfully been applied for the characterization of high-density polyethylene, isotactic polypropylene, and impact modified polypropylene copolymers. For the first time, NMR experiments on polyolefin materials have been performed at both low and high field. These NMR methods alone and in combination with DSC, X-ray, IR, TEM, and AFM were used to obtain information about the changes in the structure, molecular dynamics, and domain thickness of the investigated systems. These changes were caused by stretching and exposure of the sample at elevated temperature for a certain time, and addition of rubber and nucleating agents. The main conclusions of this work results are as follows: 1. The most appropriate description of the morphology of both homopolymer HDPE and iPP by NMR is a three-phase model (rigid, semi-rigid, and soft fractions). The amounts of the rigid, semi-rigid and soft fraction, and chain dynamics in each fraction of polyethylene and isotactic polypropylene have been investigated by 1H solid state NMR as a function of temperature. The domain thicknesses as determined by NMR were compared with those measured by small-angle X-ray scattering (SAXS) and transmission electron microscopy (TEM) and are in a good agreement. This proves the reliability of the low-field NMR methods for the determination of domain thicknesses in heterogeneous polymers like those enumerated above. Annealing at higher temperatures causes an increase in the lamellae thickness and crystallinity and a decrease in the chain mobility in all fractions. It was shown that the thickening of the crystalline domains during annealing of iPP can be described by a model based on irreversible thermodynamics. A phenomenological correlation was established between the relaxation rate of 1H transverse magnetization characteristic to the rigid fraction of iPP and the annealing temperatures. 2. The effects of deformation ratio, deformation rate, and deformation temperatures on the phase content, chain mobility, and domain thickness of homopolymeric iPP has been investigated using low-field solid state NMR. The amount of the rigid
119
fraction/degree of crystallinity obtained by NMR coincided within few percent with that obtained by IR. A decrease in the chain mobility has been observed in all fractions with increasing the drawing rate, draw ratio, and drawing temperature. The increase in the thickness of the lamellae with increasing draw ratio and drawing temperature can be explained by partial melting and recrystallization, as well as immobilization of the amorphous phase. We established correlations between the thicknesses of the rigid domains and the amount of the rigid fraction of unaxially stretched iPP, as well as between the thickness of the rigid domains and the chain mobility in the rigid fraction. 3. The effect of aging temperature and aging time on the phase composition, chain mobility and mechanical properties of iPP homopolymer has been studied in detail by
1
H low-field solid state NMR. The samples were obtained by a process involving
injection molding, crystallization from the melt, and quenching. The aging effect on different morphologies of homopolymeric iPP samples was rationalized in terms of increased restrictions of the chain mobility in the semi-rigid and soft fractions. All samples exhibited physical aging at elevated temperatures and furthermore, it was observed that the physical aging kinetics depends by the morphology of the iPP and the aging temperature. The structural changes due to aging are considered to be associated with a molecular rearrangement generally similar to a secondary crystallization which is accompanied by a decrease in segmental chain mobility of the diffusing amorphous chains. 4. Changes in the morphology, phase content, chain mobility, and mechanical data of impact modified polypropylene copolymer have been investigated using a combination of different techniques such as NMR, X-Ray, TEM, AFM, and DSC. These changes were caused by the ethylene-propylene rubber and nucleating agents added to improve the impact strength and modulus of the polymer. A decrease in the heat of fusion, yield stress, flexural modulus has been observed with increasing rubber content. The addition of nucleating agents however, caused an increase in the crystallization temperature, heat of fusion, yield stress, flexural modulus, and rigid fraction. Based on the above results, we can conclude that low and high field proton NMR can be successfully used to obtain information about the structure, molecular dynamics, and domain 120
thickness in the investigated homopolymers of HDPE, iPP, and copolymeric PP. However, the potential of these NMR techniques can be easily extended to other polymeric systems. Corroborating the NMR results with data obtained from other measurement techniques as those enumerated above, helped in gaining further insights in the microscopic properties of the materials under investigation. Therefore, the authors hopes that this thesis brings the reader a step closer to understanding at least part of the known multiple possibilities of NMR investigations and sets a solid base for those that are yet to be discovered.
121
122
References
References
[Abr1961] [Aga1962] [Aga1981] [Aha1983] [Aki1999] [Ala1999] [Ale1969] [Ali2001] [Ass1978] [Ath1994] [Aur2000] [Aur2001] [Awa1998] [Axe1983] [Ba1998] [Bas1974] [Bal1989] [Bal2000] [Bas1973] [Bed2000] A. Abragam, Principles of Nuclear Magnetism, Oxford University Press, Oxford, 1961. M. K.Agarwal, K. E. Helf, Kolloid Z. 180, 114-118 (1962). M. Agarwal, J. M. Schultz, Polym. Eng. Sci. 21, 776 781 (1981). S. M. Aharoni, Macromolecules 16, 1722-1728 (1983). A. Akizadeh, L. Richardson, J. Xu, S. McCartney, H. Marand, Y. W. Cheung, S. Chum, Macromolecules 32, 6221 (1999). R. G. Alamo, G. M. Brown, L. Mandelkern, A. Lehtinen, R. Paukkerri,
Polymer 40, 3933 (1999).
L. E. Alexander, Diffraction Methods in Polymer Science, John Wiley & Sons, Inc.: New York, 1969. A. Alizahedh, S. Sohn, J. Quinn, H. Marand, L. C. Shank, H. D. Iler,
Macromolecules 34, 4066 (2001).
R. A. Assink, Macromolecules 11, 1233 (1978). ATHAS Data Bank. http://web.utk.edu/~athas/databank/ Ed. M. Pyda, 1994. F. Auriemma, O.R. Ballesteros, C. De Rosa, P. Corradini, Macromolecules 33, 8764, (2000). F. Auriemma, C. De Rosa, T. Boscato, P. Corradini, Macromolecules 34, 4815 (2001). H. Awaya, Polymer 26, 591, (1988). D. E. Axelson, L. Mandelkern, R. Popli, P.J. Mathieu, Polym. Sci.:
Polym.Phys. 2, 2319 (1983).
Y. Ba, J. A. Ripmesster, J. Chem. Phys. 108, 8589-8594 (1998). D. C. Bassett, S. Block, G. J. Piermarini, J. Appl. Phys. 45, 4146, (1974) F. J. Balta-Calleja, C. G. Vonk, X-ray Scattering of Synthetic Polymers
Elsevier, New York, 1989.
S. Balijepalli, G. C. Rutledge, Comp. And Theoretical Pol. Sci. 10, 103-113 (2000). D. C. Bassett, D. R. Carder, Polymer 24, 387 (1973). E.L. Bedia, N. Astrini, A. Sudarisman, F. Sumera and Y. Kashiro. J Appl
123
References
Polym Sci 78, 1200 (2000).
[Bed2002] [Bell1993] [Ber1967] [Ber1978] [Boy1979] [Boy1983] [Boy1984] [Boy1985] [Bra1986] [Bra1999] [Bri1993] [Bro1997] [Bud2003] [Bud2003] [Bud2004] [Bun1939] [Bus1990] [Cam1993] [Cla1993] [Cha1977] [Cha1980] [Cha2002] [Cha2002] [Che1994] [Che2006]
E.L. Bedia, S. Murakami, K. Senoo and S. Kohjiya. Polymer 43, 749 (2002). A. Bellare, R. E. Cohen, A. S. Argon, Polymer 34, 1393 (1993). K.Bergmann, K. Nawotki, Z-Z Polym 219, 132 (1967). K. Bergmann, J.Polym.Sci., Polym. Phys. 16, 1611 (1978). R.H. Boyd, Polym. Engng. Sci. 19, 1010 (1979). R.H. Boyd, J. Polym. Sci. Polym. Phys. 21, 493 (1983). R. H. Boyd, Macromolecules 17, 903 (1984). R. H. Boyd, Polymer 26, 323 (1985). J. M. Brady, E. L. Thomas, Polymer Prep. 27, 1 (1986). J. Brandrup, E. H. Immergut, E. A. Grulke, Polymer Handbook 1999. J. V. McBrierty, K. J. Paker, Nuclear Magnetic Resonance in Solid Polymers
Cambridge Univ. Press, Cambridge, 1993.
G.M. Brown and J.H. Butler. Polymer 38, 3937 (1997). A. Buda, D. E. Demco, M. Bertmer, B. Blmich, V. M. Litvinov, J. P. Penning, J. Phys. Chem. B 107 5357 (2003). A. Buda, D. E. Demco, M. Bertmer, B. Blmich, B. Reining, H. Keul, H. Hcker, Solid State Nucl. Magn. Reson. 24, 39 2003. A. Buda, D. E. Demco, B. Blmich, V. M. Litvinov, J. P. Penning,
ChemPhysChem 5, 876-883 (2004).
C. W. Bunn, Trans. Farad. Soc. 35, 482, (1954). W. R. Busing, Macromolecules 23, 4608, (1990). R. A. Campbell, P. J. Philips, Polymer 34, 4809, (1939). J. Clauss, K. Schmidt-Rohr, H. W. Spiess, Acta Polym. 44, 1-17 (1993). Y. Chantani, Y. Ueda, H. Tadokoro, Annual meeting of the Society of Polymer
Science, Japan, Tokyo, Preprint, p 1326, (1977).
C. K. Chai, N. G. McCrum, Polymer 21, 706 (1980). A.C. Chang, L. Tau, A. Hiltner and E. Baer. Polymer 43, 4923 (2002). A.C. Chang, S.P. Chum, A. Hiltner and E. Baer. J Appl Polym Sci 86, 3625 (2002). J. Cheng, M. Fone, V. N. Reddy, K. B. Schwartz, H. P. Fisher, B. Wunderlich
J.Polym.Sci.: Part B: Polym.Phys. 32, 2683 (1994).
124
References [Cho1988] [Chi1961] [Cop1987] [Cru1984] [Cru2004] [Dad1994] [Dem1995] [Dou1979] [Ear1979] [Eck1997] [Ern1987] J.C. Chou, K. Vijayan, D. Kirby, A. Hiltner, E. Baer, J. Mater. Sci. 23, 2521 (1988). R. Chiang, P.J. Flory, J. Am. Chem. Soc. 83, 2857 (1961). F. Coppola, R. Greco, E. Martucelli, H.W. Kammer and C. Kummerlowe.
Polymer 28, 47 (1987).
N. G. McCrum, Polymer 25, 299 (1984). N. G. McCrum, B. E. Read, G. Williams, Anelastic and Dielectric Effects in Polymeric Solids, Wiley, New York, 2004. D. Dadayli, R. K. Harris, A. M. Kenwright, B. J. Say, M. M. Sunnetcioglu,
IRC Polym. and Tech. 35, 4083 (1994).
D. E. Demco, A. Johansson, J. Tegenfeldt, Solid State Nucl. Magn. Reson. 4, 13-38 (1995). D.C. Douglass.; V.J. McBrierty, Polym. Eng. Sci. 19, 1054, (1979) and references therein. W. Earl, D. L. VanderHart, Macromolecules 12, 762 (1979). R. R. Eckman, P. M. Henrichs, A. J. Peacock, Macromolecules 30, 2474 (1997). R. R. Ernst, G. Bodenhausen, A. Wokaun Principles of Nuclear Magnetic Resonance in One and Two Dimensions, Oxford University Press, Clarendon, 1987.
[Fed1985] [Fer1984] [Fil1994] [Fis1962] [Fis1969] [Flo1978] [Fuj1991] [Gau2000] [Ghe1994] [Goi1992] [Goi1993] [Gol1961]
V. D. Fedotov, N. A. Abdrashitiva. Polym. Sci. USSR 27, 287 (1985). A. Ferrero, E. Ferracini, R. Hosemann, Polymer 25, 1747 (1984). B. Fillon, A. Thierry, B. Lotz, J. Wittman, J. Therm. Analys. 42, 721 (1994) E. W. Fischer, G. F. Z. Schmidt, Angew. Chem. 1, 488 (1962). E. W. Fischer, M. Goddar, J. Polymer Sci. C16, 4405 (1969). P. J. Flory, D. Y. Yoon, Nature 272, 226-229 (1978). M. Fujiyama, T. Walino, J. Appl. Polym. Sci. 42, 2739 (1991). S. Gautman, S. Balijepalli, G. C. Rutledge, Macromolecules 33, 9136-9145 (2000). J. Gheng, M. Fone, N. V. Reddy, B. K. Schwartz, P. H. Fisher, B. Wunderlich,
J. Polym. Sci.: Polym. Phys. 32, 2683-2693 (1994).
G. Goizueta, T. Chiba and T. Inoue, Polymer 33, 886 (1992). G. Goizueta, T. Chiba and T. Inoue, Polymer 34, 253 (1993). M. Goldman, L. Shen, Phys. Rev. 144, 321 (1961). 125
References [Gos1996] [Gre2002] [Hai1965] [Hal1972] [Han1974] [Han1998] [Hed2007] [Hed2007] [Hei1959] [Hel1997] [Hen1984] [Hik1997] [Hill1998] [Hir1990] [His1999] [Hu1999] [Hu2000] [Hu2000] [Ibh1996] [Idi1996] [Iij2000] [Isa1999] [Ito1981] [Ito1992] U. Goschel, K. Deutscher, V. Abetz, Polymer 37, 1 (1996). C. Grein, C.J.G. Plummer, H.-H. Kausch, Y. Germain and P. Beguelin.
Polymer 43, 3279 (2002).
I. L. Hay, A. Keller, Kolloid-Z. & Z. Polymere 204, 43 (1965). J. C. Halpin, J. L. Kardos, J Appl. Phys. 43, 2235 (1972). J. M. Hannon, J. Appl. Poly. Sci. 18, 3761, (1974). E. W. Hansen, P. E. Kristiansen , B. Pedersen, J.Phys..Chem. B 102, 5444 (1998). C. Hedesiu, D. E. Demco, R. Kleppinger, A. A. Buda, B. Blumich, K. Remerie, V. M. Litvinov, Polymer 48, 763 (2007). C. Hedesiu, D. E. Demco, R. Kleppinger, G. Vanden Poel, W. Gijsbers, B. Blmich, K. Remerie, V. M. Litvinov, Macromolecules, 40, 3977-3989 (2007). W. Heinsen, Journal of Polymer Science 38, 134 1959. S. Hellinckx , Colloid Polym. Sci. 275, 116 (1997) D. Hentschel, H. Sillescu, H. W. Spiess, Polymer 25, 1078 (1984). M. Hikosaka, K. Amano, S. Rastogi, A. Keller, Macromolecules 30, 2067 (1997). L. Hillebrand, A. Schmidt, A. Bolz, M. Hess, W. Veeman, R. J. Meier, G. Van der Velden, Macromolecules 31, 5010 (1998). J. Hirschinger, H. Miura, A. D. English, Macromolecules 23, 2153 (1990). R. Hiss, S. Hobeika, C. Lynn, G. Strobl, Macromolecules 32, 4390 (1999). W. G. Hu, C. Boeffel, K. Schmidt-Rohr, Macromolecules 32, 1611-1619 (1999). W.-G. Hu, K. Schmidt-Rohr, Polymer 41, 2979 (2000). J. Z. Hu, W. Wang, S. Bai, R. J. Pugmire, G. M..V. Taylor, D. M. Grant,
Macromolecules 33, 3359 (2000).
A. O. Ibhadon, J. Appl. Polym. Sci. 62, 1843 (1996). D. S. H. Idiyatullin, E. V. Khozina, V. S. Smirnov, Solid State Nucl. Magn.
Reson. 7, 17 (1996).
M. Iijima, G. Strobl, Macromolecules 33, 5204 (2000). J. R. L. Isasi, M. J. Mandelkern, R. G. Galante Alamo, J. Polym. Sci., Polym.
Phys. 37, 323 (1999).
M. Ito, T. Kanamoto, K. Tanaka, Macromolecules 14, 1779 (1981). J. Ito, K. Mitani, Y. Mizutani, J. Appl. Polym. Sci. 46, 1221 (1992). 126
References [Iwa2002] [Jan1993] [Jan1994] [Jan1996] [Kak1985] [Kar1979] [Kar1995] [Kar1999] [Kav1971] [Kha1996] [Kim1996] [Kim1998] [Kis1983] [Kit1977] [Kit1978] [Kit1986] [Kit1994] [Kiv1986] [Kle2002] [Kri1999] [Kri2000] [Kuw2000] [Lab2001] [Li1999] [Lin1974] [Lit2002] Iwata K. Polymer 43, 6609 (2002). J. Jancar, A. T. J. Dibenedetto, J. Kucera, Polymer 34, 1684 (1993). J. Jancar, A. T. J. Dibenedetto, J.Matter. Sci. 29, 4651 (1994). J. Jancar, J.Matter. Sci. 31, 3983 (1996). T. Kakudate, M. Kakizaki , T. Hideshima, J.Polym.Sci.: Polym.Phys. 23, 787 (1985). J. Karger-Kocsis, A. Kallo, A. Szafner, G. Bodar and Z. Senyei. Polymer 20, 37 (1979). J. Karger Kocsis, Polypropylene; structure, blends and composites; Vol. 1, Champman and hall, London, (1995). H. G. Karian, Handbook of Polypropylene and Polypropylene Composites; Marcel Dekker: New York, 1999. S. Kavesh, J.M. Schultz, J. Polym. Sci. Part A-2 9, 85 (1971). A. Khare, A. Mitra, J. Mater. Sci. 31, 5691 (1996). G.-M. Kim, G.H. Michler, M. Gahleitner and J. Fiebig. J Appl Polym Sci 60, 1391 (1996). G.-M. Kim and G.-H. Michler. Polymer 39, 5689 (1998). Y. V. Kissin, J. of Polym. Sci.: Polym. Phys. Edit. 31, 2085 (1983). R. Kitamaru, F. Horii, F. Hyon, J Polym Sci Polym Phys 15, 821 (1977). R. Kitamaru, F. Horii, Adv Polym. Sci. 26, 139 (1978). R. Kitamaru, F. Horii, K. Murayama, Macromolecules 19, 636 (1986). R. Kitamaru, F. Horii, Q. Zhu, D. C. Bassett, R. H. Olley, Polymer 35, 1171 (1994). A. Kivinen, M. Ovaska, K. Soljamo, Chemical Physics Letters 128, 4 (1986). P. G. Klein, M. A. N. Driver, Macromolecules 35, 6598 (2002). P. E. Kristiansen, E. W. Hansen, B. Pedersen, J.Phys..Chem. B: 103, 3552 (1999). P. E. Kristiansen, E. W. Hansen, B. Pedersen, Polymer 41, 311 (2000). K. Kuwabara, H. Kaji, M. Tsuji , F. Horii, Macromolecules 33, 7093 (2000). T. Labour, C. Gauthier, R. Seguela, G. Vigier, Y. Bomal, G. Orange, Polymer
42, 7127 (2001).
J. X. Li, W. L. Cheung, C. M. Chan, Polymer 40, 2089 (1999). L. Lin, A. S. Argon, J. Mater. Sci. 9, 2034 (1974). V.M. Litvinov, V. B. F. Mathot, Solid State Nuclear Mag. Res. 22, 218 127
References (2002). [Lit2004] [Lit2005] [Lot1991] [Lot1996] [Lov1983] [Mag1996] [Mai2000] [Man1964] [Man1968] [Man1985] [Man1988] [Man1990] [Man1990] [Mar2000] [Mar1998] [Mat1994] [Mat2000] [Men2003] [Mez1994] [Mez1997] [Mes1995] [Mun1987] [Nat1955] [Nat1956] V.M. Litvinov, J. P. Penning, Macromol. Chem. Phys. 205, 172 (2004). V. M. Litvinov, M. Soliman, Polymer 46, 3077 (2005). B. Lotz, S. Graff, C. Staupe, J. C. Wittman, Polymer 32, 2903, (1991). B. Lotz, J. C. Wittmann, Polymer 37, 4979, (1996). A. J. Lovinger, J. Polm. Phys. 21, 97, (1983). S. N. Magonov, Surface analysis with STM and AFM, VCH Publisher, New York, 1996 P. Maiti, M. Hikosaka, K. Yamada, A. Toda, F. Gu, Macromolecules 33, 9069 (2000). L. Mandelkern, Crystallization of polymers, McGraw-Hill, New York, 1964. L. Mandelkern, A. L. Allou, M. Gopalan, J. Phys. Chem. 72, 309 (1968). L. Mandelkern, Polymer J. 17, 337 (1985). L. Mandelkern, A. J. Peacock, Stud. Phys. Theor. Chem. 54, 201-217 (1988). L. Mandelkeren, R. G. Aloma, M. A. Kennedy, Macromolecules 23, 47214723 (1990). L. Mandelkern, Acc. Chem. Res. 23, 380-394, (1990). H. Marand, A. Alizadeh, R. Farmer, R. Desai, V. Velikov, Macromolecules
33, 3392 (2000).
A. Martorana, S. Piccarolo, D. Sapoundjieva, Macromolec. Chem. Phys. 3, 531 (1998). V. B. F. Mathot, Calorimetry and Thermal Analysis of Polymers Hanser
Publishers: Munich, Germany, Vienna, and New York, 1994.
H. Matsuda, T. Aoike, H. Uehara, T. Yamanobe, T. Komoto, Polymer 42, 5013-5021 (2000). R. Mendoza, G. Regnifer, W. Seiler, J. L. Lebrun, Polymer 44, 3363 (2003). K. Mezghani, R. A. Cambell, P. J. Phillips, Macromolecules 27, 997 (1994). K. Mezghani, P. J. Phillips, Polymer 38, 5725, (1997) P. B. Messersmith, E. P. Giannelis, J. Polym. Sci. Part A: Polym. Chem. 33, 1047 (1995). M. Munowitz, A. Pines, Adv. Chem. Phys. 66, 1-152 (1987). G. Natta, P. Pino, P. Corrandini, F. Danusso, E. Mantica, G. Mazzanti, G. Moragilo, J. Am. Chem. Soc. 77, 1708, (1955). G. Natta, Angew. Chem. 68, 541, (1956). 128
References [Nat1964] [Neg1992] [Nor1995] [Nys1995] [Nys1999] [Oll1989] [Ora1982] [Ora2001] [Pac1984] [Pau1993] [Pav2005] [Pea1990] [Per1999] [Pet1963] [Pet1967] [Pet1971] [Pet1973] [Pet1987] [Pet1996] [Poe2003] [Pop1987] [Pou1998] [Pre2000] [Ros1992] [Ros2002] [Rul1961] G. Natta, Angew. Chem. 76, 553, (1964). M. Negahban, A. S. Wineman, Int. J. Engng. Sci. 30, 819 (1992). T. Normura, T. Nishio, T. Fujii, J. Sakai, M. Yamamoto, A. Uemura, M. Kakugo. Polym Engng Sci. 35, 1261, 1995. B. Nysten, R. Legras and J.-L. Costa. J. Appl. Phys. 78, 5953 (1995). B. Nysten, A. Ghanem, J.-L. Costa and R. Legras. Polym. Int. 48, 334 (1999). R. H. Olley, D. C. Bassett, Polymer 30, 399-409 (1989). L. D'Orazio, N. Greco, C. Mancarella, E. Martuscelli, G. Nagosta and C. Silvereste. Polym Engng Sci 22, 536 (1982). L. D'Orazio and G. Cecchin. Polymer 42, 2675 (2001). K. J. Packer, J. M. Pope, R. R. Yeung, J. Polym. Sci.: Polym. Phys. 22, 589616 (1984). R. Paukkeri, T. Vaananen, A. Lehtinen, Polymer 34, 2488, (1993). S. Pavel, T. Hashimoto, S. Kenji, G. Arkadiusz, Polymer 46, 513-521, (2005). A. Peacock, L. Mandelkern, J. Polym. Sci., Polym. Phys. Ed. 28, 1917, (1990). E. Perez, D. Zucchi, M. C. Sacchi, F. Forlini, A. Bello, Polymer 40, 675, (1999) A. Peterlin, J. Polym. Sci. B1 B1, 279 (1963). A. Peterlin, Polymer 6, 25 (1967). A. Peterlin, J. Mater. Sci. 6, 490 (1971). A. Peterlin, Appl. Polym. Symp. 20, 269 (1973). A. Peterlin, Colloid&Polymer Sci. 265, 357 (1987). F. Coppola, R. Greco, E. Martucelli, H.W. Kammer and C. Kummerlowe.
Polymer 28, 47 (1987).
R. Popli, L. Mandelkern, J. Polym. Sci. , Polym. Phys. 25, 441-449, (1987). L. Poussin, Y. A. Bertin, J. Parisot, C. Brassy, Polymer 39, 4261 (1998). R. Premphet, P. Horanont, J. Appl. Polym. Sci. 76, 1929 (2000). S. Rostami, I. S. Miles, Multicomponent Polymer Systems Wiley: New York, 1992. C. De Rosa, F. Auriemma, T. Circelli, R. M. Waymouth, Macromolecules 35, 3622, (2002). W. Ruland, Acta Crystallogr. 14, 1180 (1961). 129
References [Rul1964] [Sam1981] [San1973] [San1974] [San1986] [Set1968] [Son1987] [Sch1994] [Sch1991] [Sch1999] [Sch1999] [Sil1998] [Sta1961] [Sta1998] [Ste1969] [Str1980] [Str1997] [Str1987] [Swa1998] [Tan1989] [Tan1993] [Tea1957] [Tho1996] [Tom1996] [Tom1998] [Tom1998] [Tom2000] W. Ruland, Polymer 5, 89 (1964). R. J. Samuels, Makromol. Chem. 4, 241 (1981). I. C. Sanchez, J. P. Colson , R. K. Egy, J. Appl. Phys. 44, 4332 (1973). I. C. Sanchez, A. Peterlin, R. K. Egy, F. L. McCrackin, J. Appl. Phys. 45, 4216 (1974). H. Sano, T. Usami and H. Nakagawa, Polymer 27, 1497 (1986). T. Seto, T. Hara, K. Tanaka, Japanese J. Phys. 48, 282, (1968) H. H. Song, R. J. Roe, Macromolecules 20, 2723 (1987). K. Schmidt-Rohr, H. W. Spiess, Multidimensional Solid-State NMR and Polymers , Academic Press, London, 1994. K. Schmidt-Rohr, H. W. Spiess, Macromolecules 24, 5288 (1991). M. Schneider, L. Gasper, D. E. Demco, B. Blmich, J. Chem. Phys. 111, 402 (1999). S. Schreurs, J. P. Francois, P. Adriaenens, J. Gelan, J.Phys.Chem. B 103, 1393 (1999). R. Silvestry, P. Sgarzi, Polymer 39, 5871 (1998). W. C. Statton, J. Appl. Phys. 32, 2332 (1961). J.U. Starke, G.U. Michler, W. Grellmann, S. Seidler, M. Gahleitner, J. Fiebig and E. Nezbedova. Polymer 39, 75 (1998). R. S. Stein, Polymer Eng. Sci. 9, 320 (1969). G. R. Strobl, M. J. Schneider, I. G. Voigt-Martin, J.Polym.Sci. 18, 1361 (1980). G. R. Strobl, The Physics of Polymers, Springer-Verlag, Berlin 1997. L. C. E. Struik, Polymer 28, 1521 (1987). K. Swaminathan and D.W.M. Maw. J Appl Polym Sci 78, 452 (2000). H. Tanaka, F. Kohrogi, K. Suzuki, Eur. Polym. J. 25, 449 (1989). H. Tanaka, Y. Inoue, Polymer International 31, 9 (1993). P. W. Teare, D. R. Holmes, J. Polym. Sci. 24, 496, (1957). R. Thomann, C. Wang, J. Kressler, Macromolecules 29, 8425, (1996). P. E. Tomlins, Polymer 37, 3907 (1996). P. E. Tomlins, B. E. Read, Polymer 39, 355 (1998). E. Tomasetti, B. Nysten and R. Legras. Nanotechnology 9, 305 (1998). E. Tomasetti, R. Legras, B. Henri-Mazeaud and B. Nysten. Polymer 41, 6597 (2000). 130
References [Tri1998] [Ueh2000] [Uzo1997] [Van1996] [Van2003] [Var1992] [Vle1997] [Vod2006] [Von1971] [Von1973] [Vor1996] [Wan1996] [Web1] [Web2] [Weg2000] [Wen1993] [Woo1993] [Wun2005] [Yam1998] [Yam1999] [Yam2003] [Yok1997] [Zie1955] [Zha2000] [Zha2002] [Zho2005] D. Trifonova, J. Varga, G.J. Vancso, Polm. Bull. 41, 341, (1998). H. Uehara, T. Yamanobe, (2000). T. C. Uzomah, S. C. O Ugbolue, J Appl Polym. Sci. 65, 625 (1997). D. L. VanderHart, G. B. McFadden, Solid State Nucl. Magn. Reson. 7, 45-66 (1996). D.L.VanderHart, C. R. Snyder, Macrololecules 36, 4813-4826 (2003). J. Varga, J. Mat. Sci. 27, 2557, (1992). S. Vleeshouwers, Polymer 38, 3213, (1997) A. M. Voda, D. E. Demco, A. Voda, T. M. SchauberAdler, T. Dabisch, A. Adams, M. Baias, B. Blmich, Macromolecules, 39, 4802 (2006). C. G. Vonk, J.Appl.Crystallogr 4, 340 (1971). C. G. Vonk, J.Appl.Crystallogr 6, 149 (1973). F. J. Wortmann, K.V. Schulz, Polymer 37, 819 (1996). J. Wang, J.Chem.Phys. 104, 48 (1996). www.yellow-design.de www.imc.cas.cz W. P. Weglarz, H. Peemoeller, A. Rudin, J Polym Sci, B: Polym Phys. 38, 2487 (2000). W. Wenig, F. Herzog, J. App. Polym. Sci. 50, 2163 (1993). R. P. Wool, Macromolecules 26, 1564 (1993). B. Wunderlich, Thermal analysis of polymeric materials, Springer, 2005. K. Yamada, S. Matsumoto, K. Tagashira, M. Hikosaka Polymer 39, 5327 (1998). M. Yamaguchi and K.H. Nitta. Polym Engng Sci 39, 833 (1999). K. Yamada, M. Hikosaka, A. Toda, S. Yamazaki, K. Tagashira
Macromolecules 36, 4790 (2003).
T.
Y. Yokoyama and T. Ricco. J Appl Polym Sci 66, 1007, (1997). K. Ziegler, E. Holzkamp, H. Briel, H. Martin, Angew. Chem. 67, 426, (1955). X. C. Zhang, M. F. Butler, R. E. Cameron, Polymer 41, 3797 (2000). M. Zhang, Y. Liu, X. Zhang, J. Gao, F. Huang, Z. Song, G. Wei and J. Qiao.
Polymer 43, 5133 (2002).
131
References [Zhu2004] [Zhu2005] P. Zhu, E. Graham, Polymer 45, 2603 (2004). P. Zhu, J. Tung, G. Edward, Polymer 46, 10960 (2005).
132
Acknowledgements
Acknowledgements
At the end of this thesis I would like to express my gratitude to everyone who directly or indirectly contributed to this work and its completion. First and foremost I must single out my promoter Prof. Dr. Dr. h. c. Bernhard Blmich who gave me the opportunity to perform my PhD in his group. It was a pleasure and a privilege for me to be part of his team. Many thanks to Prof. Dan Demco for his permanent advices and help related to any NMR problem. I would kindly like to thanks Dr. Klaas Remerie, Dr. Rieky Steenbakkers, Dr Maria
Soliman from SABIC Europe, for many hours of useful discussions, for the support and
understanding during these years. Thank you for being around anytime I needed. Many thanks are addressed to Dr. Victor Litvinov, Dr. Ralf Kleppinger, Dr. Rudy Debliek from DSM Research Resolve for their support and interesting discussions we had. I would like to thank my colleagues and friends from RWTH Aachen Dr. Alina Adams, Dr.
Adrian Voda, Lavinia Utiu, Maria Baias, Isa Fonseca, Elena Talnishnikh, Claudiu Melian for good advices and nice time we had together.
I would like to thank my colleagues and friends from DSM Research and SABIC Europe Dr.
Mihai Morariu, Dr. Elena Trezza, Harrie Linssen, Gerda Kolfschoten and Dr. Aurora Bastinas for the nice moments we have shared together during the coffee breaks and for the
friendly environment. For sure I do not forget all my friends from the five stars hotel AMVJ Geleen where I had very nice time during my PhD thesis. At the end I want to say the biggest thanks for my mother Valeria and my father Virgil who gave me the power to go on. I am also deeply thankful to my sister Ioana and my brother
Nicu.
133
Curriculum Vitae
Curriculum Vitae
Cristian Eugen Hedesiu
Birth date and place: 08 April 1980, Iclod, Romania E-mail: cristihed@gmail.com Nationality: Romanian
Education
2007 December Thesis: 04. 2004 2006 09.2002 08.2003 Dissertation: 09.1998 07.2002 Dipolma:
Defence for the Doctoral Degree in natural science Structure Property Relationship in Semi - Crystalline Polymers (Polyethylene and Polypropylene)
PhD Student, RWTH Aachen, Germany and SABIC Europe B.V., The Netherlands
Master of science (Biophysics and Medical Physics), BabesBolyai University, Cluj-Napoca, Romania, in collaboration with RWTH, Aachen, Germany Investigation of Nylon6 fibers by Solid State NMR Dip. Phys., Faculty of Physics, Babes-Bolyai University, ClujNapoca, Romania Application of gamma rays in cardiology
Professional Experience
2004 2007 2003 2002
SABIC Research Campus, Geleen, The Netherlands; DSM Resolve Research Campus, Geleen, The Netherlands; IIT Research Institute, New Delhi, India. DAAD Program; Erasmus Socrates Scholarship at the Institute for Technical and Macromolecular Chemistry, RWTH Aachen, Germany; Lecturer at Emil Racovita High-school, Cluj, Romania; N.S. Stancioiu Heart Institute, Cluj-Napoca, Romania Nuclear Cardiology Department;
Publications
1. C. Hedesiu, D. E. Demco, R. Kleppinger, A. A. Buda, B. Blmich, K. Remerie, V. Litvinov, The effect of temperature and annealing on the phase composition, molecular mobility and the thickness of domains in high-density polyethylene, Polymer, 48 (2007) 763-777 2. C. Hedesiu, D. E. Demco, R. Kleppinger, G. Vanden Poel, W. Gijsbers, B. Blmich, K. Remerie, V. Litvinov, The effect of temperature and annealing on the phase composition, molecular mobility and the thickness of domains in isotactic
134
Curriculum Vitae
polypropylene studied by proton solid-state NMR, SAXS and DSC, Macromolecules, 40 (2007) 3977 3989. 3. C. Hedesiu, D. E. Demco, B. Blmich, K. Remerie, V.M Litvinov, Study of uniaxially deformed isotactic polypropylene by 1H solid state NMR and IR spectroscopy. Accepted to Macromolecular Chemistry and Physics
4. C. Hedesiu, D. E. Demco, R. Steenbakkers, R. Kleppinger, G. Vanden Poel, B. Blmich, K. Remerie, V.M Litvinov, Aging Study on Polypropylene by 1H solid state NMR, X-ray, and DSC. In preparation 5. C. Hedesiu, D. E. Demco, R. Steenbakkers, B. Blmich, K. Remerie, V.M Litvinov, Study of PP impact polypropylene by 1H solid state NMR, SAXS, TEM and SEM. In preparation
135